20210608.md

ArXiv cs.CV --Tue, 8 Jun 2021

1.SIMONe: View-Invariant, Temporally-Abstracted Object Representations via Unsupervised Video Decomposition ⬇️

To help agents reason about scenes in terms of their building blocks, we wish to extract the compositional structure of any given scene (in particular, the configuration and characteristics of objects comprising the scene). This problem is especially difficult when scene structure needs to be inferred while also estimating the agent's location/viewpoint, as the two variables jointly give rise to the agent's observations. We present an unsupervised variational approach to this problem. Leveraging the shared structure that exists across different scenes, our model learns to infer two sets of latent representations from RGB video input alone: a set of "object" latents, corresponding to the time-invariant, object-level contents of the scene, as well as a set of "frame" latents, corresponding to global time-varying elements such as viewpoint. This factorization of latents allows our model, SIMONe, to represent object attributes in an allocentric manner which does not depend on viewpoint. Moreover, it allows us to disentangle object dynamics and summarize their trajectories as time-abstracted, view-invariant, per-object properties. We demonstrate these capabilities, as well as the model's performance in terms of view synthesis and instance segmentation, across three procedurally generated video datasets.

2.Mean-Shifted Contrastive Loss for Anomaly Detection ⬇️

Deep anomaly detection methods learn representations that separate between normal and anomalous samples. Very effective representations are obtained when powerful externally trained feature extractors (e.g. ResNets pre-trained on ImageNet) are fine-tuned on the training data which consists of normal samples and no anomalies. However, this is a difficult task that can suffer from catastrophic collapse, i.e. it is prone to learning trivial and non-specific features. In this paper, we propose a new loss function which can overcome failure modes of both center-loss and contrastive-loss methods. Furthermore, we combine it with a confidence-invariant angular center loss, which replaces the Euclidean distance used in previous work, that was sensitive to prediction confidence. Our improvements yield a new anomaly detection approach, based on $\textit{Mean-Shifted Contrastive Loss}$, which is both more accurate and less sensitive to catastrophic collapse than previous methods. Our method achieves state-of-the-art anomaly detection performance on multiple benchmarks including $97.5%$ ROC-AUC on the CIFAR-10 dataset.

3.NTIRE 2021 Challenge on Burst Super-Resolution: Methods and Results ⬇️

This paper reviews the NTIRE2021 challenge on burst super-resolution. Given a RAW noisy burst as input, the task in the challenge was to generate a clean RGB image with 4 times higher resolution. The challenge contained two tracks; Track 1 evaluating on synthetically generated data, and Track 2 using real-world bursts from mobile camera. In the final testing phase, 6 teams submitted results using a diverse set of solutions. The top-performing methods set a new state-of-the-art for the burst super-resolution task.

4.High Resolution Solar Image Generation using Generative Adversarial Networks ⬇️

We applied Deep Learning algorithm known as Generative Adversarial Networks (GANs) to perform solar image-to-image translation. That is, from Solar Dynamics Observatory (SDO)/Helioseismic and Magnetic Imager(HMI) line of sight magnetogram images to SDO/Atmospheric Imaging Assembly(AIA) 0304-Å images. The Ultraviolet(UV)/Extreme Ultraviolet(EUV) observations like the SDO/AIA0304-Å images were only made available to scientists in the late 1990s even though the magenetic field observations like the SDO/HMI have been available since the 1970s. Therefore by leveraging Deep Learning algorithms like GANs we can give scientists access to complete datasets for analysis. For generating high resolution solar images we use the Pix2PixHD and Pix2Pix algorithms. The Pix2PixHD algorithm was specifically designed for high resolution image generation tasks, and the Pix2Pix algorithm is by far the most widely used image to image translation algorithm. For training and testing we used the data for the year 2012, 2013 and 2014. The results show that our deep learning models are capable of generating high resolution(1024 x 1024 pixels) AIA0304 images from HMI magnetograms. Specifically, the pixel-to-pixel Pearson Correlation Coefficient of the images generated by Pix2PixHD and original images is as high as 0.99. The number is 0.962 if Pix2Pix is used to generate images. The results we get for our Pix2PixHD model is better than the results obtained by previous works done by others to generate AIA0304 images. Thus, we can use these models to generate AIA0304 images when the AIA0304 data is not available which can be used for understanding space weather and giving researchers the capability to predict solar events such as Solar Flares and Coronal Mass Ejections. As far as we know, our work is the first attempt to leverage Pix2PixHD algorithm for SDO/HMI to SDO/AIA0304 image-to-image translation.

5.3DB: A Framework for Debugging Computer Vision Models ⬇️

We introduce 3DB: an extendable, unified framework for testing and debugging vision models using photorealistic simulation. We demonstrate, through a wide range of use cases, that 3DB allows users to discover vulnerabilities in computer vision systems and gain insights into how models make decisions. 3DB captures and generalizes many robustness analyses from prior work, and enables one to study their interplay. Finally, we find that the insights generated by the system transfer to the physical world.
We are releasing 3DB as a library (this https URL) alongside a set of example analyses, guides, and documentation: this https URL .

6.Visual Transformer for Task-aware Active Learning ⬇️

Pool-based sampling in active learning (AL) represents a key framework for an-notating informative data when dealing with deep learning models. In this paper, we present a novel pipeline for pool-based Active Learning. Unlike most previous works, our method exploits accessible unlabelled examples during training to estimate their co-relation with the labelled examples. Another contribution of this paper is to adapt Visual Transformer as a sampler in the AL pipeline. Visual Transformer models non-local visual concept dependency between labelled and unlabelled examples, which is crucial to identifying the influencing unlabelled examples. Also, compared to existing methods where the learner and the sampler are trained in a multi-stage manner, we propose to train them in a task-aware jointly manner which enables transforming the latent space into two separate tasks: one that classifies the labelled examples; the other that distinguishes the labelling direction. We evaluated our work on four different challenging benchmarks of classification and detection tasks viz. CIFAR10, CIFAR100,FashionMNIST, RaFD, and Pascal VOC 2007. Our extensive empirical and qualitative evaluations demonstrate the superiority of our method compared to the existing methods. Code available: this https URL

7.DoubleField: Bridging the Neural Surface and Radiance Fields for High-fidelity Human Rendering ⬇️

We introduce DoubleField, a novel representation combining the merits of both surface field and radiance field for high-fidelity human rendering. Within DoubleField, the surface field and radiance field are associated together by a shared feature embedding and a surface-guided sampling strategy. In this way, DoubleField has a continuous but disentangled learning space for geometry and appearance modeling, which supports fast training, inference, and finetuning. To achieve high-fidelity free-viewpoint rendering, DoubleField is further augmented to leverage ultra-high-resolution inputs, where a view-to-view transformer and a transfer learning scheme are introduced for more efficient learning and finetuning from sparse-view inputs at original resolutions. The efficacy of DoubleField is validated by the quantitative evaluations on several datasets and the qualitative results in a real-world sparse multi-view system, showing its superior capability for photo-realistic free-viewpoint human rendering. For code and demo video, please refer to our project page: this http URL.

8.Drone-based AI and 3D Reconstruction for Digital Twin Augmentation ⬇️

Digital Twin is an emerging technology at the forefront of Industry 4.0, with the ultimate goal of combining the physical space and the virtual space. To date, the Digital Twin concept has been applied in many engineering fields, providing useful insights in the areas of engineering design, manufacturing, automation, and construction industry. While the nexus of various technologies opens up new opportunities with Digital Twin, the technology requires a framework to integrate the different technologies, such as the Building Information Model used in the Building and Construction industry. In this work, an Information Fusion framework is proposed to seamlessly fuse heterogeneous components in a Digital Twin framework from the variety of technologies involved. This study aims to augment Digital Twin in buildings with the use of AI and 3D reconstruction empowered by unmanned aviation vehicles. We proposed a drone-based Digital Twin augmentation framework with reusable and customisable components. A proof of concept is also developed, and extensive evaluation is conducted for 3D reconstruction and applications of AI for defect detection.

9.CDN-MEDAL: Two-stage Density and Difference Approximation Framework for Motion Analysis ⬇️

Background modeling is a promising research area in video analysis with a variety of video surveillance applications. Recent years have witnessed the proliferation of deep neural networks via effective learning-based approaches in motion analysis. However, these techniques only provide a limited description of the observed scenes' insufficient properties where a single-valued mapping is learned to approximate the temporal conditional averages of the target background. On the other hand, statistical learning in imagery domains has become one of the most prevalent approaches with high adaptation to dynamic context transformation, notably Gaussian Mixture Models, combined with a foreground extraction step. In this work, we propose a novel, two-stage method of change detection with two convolutional neural networks. The first architecture is grounded on the unsupervised Gaussian mixtures statistical learning to describe the scenes' salient features. The second one implements a light-weight pipeline of foreground detection. Our two-stage framework contains approximately 3.5K parameters in total but still maintains rapid convergence to intricate motion patterns. Our experiments on publicly available datasets show that our proposed networks are not only capable of generalizing regions of moving objects in unseen cases with promising results but also are competitive in performance efficiency and effectiveness regarding foreground segmentation.

10.Digital Taxonomist: Identifying Plant Species in Citizen Scientists' Photographs ⬇️

Automatic identification of plant specimens from amateur photographs could improve species range maps, thus supporting ecosystems research as well as conservation efforts. However, classifying plant specimens based on image data alone is challenging: some species exhibit large variations in visual appearance, while at the same time different species are often visually similar; additionally, species observations follow a highly imbalanced, long-tailed distribution due to differences in abundance as well as observer biases. On the other hand, most species observations are accompanied by side information about the spatial, temporal and ecological context. Moreover, biological species are not an unordered list of classes but embedded in a hierarchical taxonomic structure. We propose a machine learning model that takes into account these additional cues in a unified framework. Our Digital Taxonomist is able to identify plant species in photographs more correctly.

11.Learning Dynamics via Graph Neural Networks for Human Pose Estimation and Tracking ⬇️

Multi-person pose estimation and tracking serve as crucial steps for video understanding. Most state-of-the-art approaches rely on first estimating poses in each frame and only then implementing data association and refinement. Despite the promising results achieved, such a strategy is inevitably prone to missed detections especially in heavily-cluttered scenes, since this tracking-by-detection paradigm is, by nature, largely dependent on visual evidences that are absent in the case of occlusion. In this paper, we propose a novel online approach to learning the pose dynamics, which are independent of pose detections in current fame, and hence may serve as a robust estimation even in challenging scenarios including occlusion. Specifically, we derive this prediction of dynamics through a graph neural network~(GNN) that explicitly accounts for both spatial-temporal and visual information. It takes as input the historical pose tracklets and directly predicts the corresponding poses in the following frame for each tracklet. The predicted poses will then be aggregated with the detected poses, if any, at the same frame so as to produce the final pose, potentially recovering the occluded joints missed by the estimator. Experiments on PoseTrack 2017 and PoseTrack 2018 datasets demonstrate that the proposed method achieves results superior to the state of the art on both human pose estimation and tracking tasks.

12.Few-Shot Unsupervised Image-to-Image Translation on complex scenes ⬇️

Unsupervised image-to-image translation methods have received a lot of attention in the last few years. Multiple techniques emerged tackling the initial challenge from different perspectives. Some focus on learning as much as possible from several target style images for translations while other make use of object detection in order to produce more realistic results on content-rich scenes. In this work, we assess how a method that has initially been developed for single object translation performs on more diverse and content-rich images. Our work is based on the FUNIT[1] framework and we train it with a more diverse dataset. This helps understanding how such method behaves beyond their initial frame of application. We present a way to extend a dataset based on object detection. Moreover, we propose a way to adapt the FUNIT framework in order to leverage the power of object detection that one can see in other methods.

13.Bias Mitigation of Face Recognition Models Through Calibration ⬇️

Face recognition models suffer from bias: for example, the probability of a false positive (incorrect face match) strongly depends on sensitive attributes like ethnicity. As a result, these models may disproportionately and negatively impact minority groups when used in law enforcement. In this work, we introduce the Bias Mitigation Calibration (BMC) method, which (i) increases model accuracy (improving the state-of-the-art), (ii) produces fairly-calibrated probabilities, (iii) significantly reduces the gap in the false positive rates, and (iv) does not require knowledge of the sensitive attribute.

14.HERS Superpixels: Deep Affinity Learning for Hierarchical Entropy Rate Segmentation ⬇️

Superpixels serve as a powerful preprocessing tool in many computer vision tasks. By using superpixel representation, the number of image primitives can be largely reduced by orders of magnitudes. The majority of superpixel methods use handcrafted features, which usually do not translate well into strong adherence to object boundaries. A few recent superpixel methods have introduced deep learning into the superpixel segmentation process. However, none of these methods is able to produce superpixels in near real-time, which is crucial to the applicability of a superpixel method in practice. In this work, we propose a two-stage graph-based framework for superpixel segmentation. In the first stage, we introduce an efficient Deep Affinity Learning (DAL) network that learns pairwise pixel affinities by aggregating multi-scale information. In the second stage, we propose a highly efficient superpixel method called Hierarchical Entropy Rate Segmentation (HERS). Using the learned affinities from the first stage, HERS builds a hierarchical tree structure that can produce any number of highly adaptive superpixels instantaneously. We demonstrate, through visual and numerical experiments, the effectiveness and efficiency of our method compared to various state-of-the-art superpixel methods.

15.Efficient Training of Visual Transformers with Small-Size Datasets ⬇️

Visual Transformers (VTs) are emerging as an architectural paradigm alternative to Convolutional networks (CNNs). Differently from CNNs, VTs can capture global relations between image elements and they potentially have a larger representation capacity. However, the lack of the typical convolutional inductive bias makes these models more data-hungry than common CNNs. In fact, some local properties of the visual domain which are embedded in the CNN architectural design, in VTs should be learned from samples. In this paper, we empirically analyse different VTs, comparing their robustness in a small training-set regime, and we show that, despite having a comparable accuracy when trained on ImageNet, their performance on smaller datasets can be largely different. Moreover, we propose a self-supervised task which can extract additional information from images with only a negligible computational overhead. This task encourages the VTs to learn spatial relations within an image and makes the VT training much more robust when training data are scarce. Our task is used jointly with the standard (supervised) training and it does not depend on specific architectural choices, thus it can be easily plugged in the existing VTs. Using an extensive evaluation with different VTs and datasets, we show that our method can improve (sometimes dramatically) the final accuracy of the VTs. The code will be available upon acceptance.

16.Unsupervised Action Segmentation for Instructional Videos ⬇️

In this paper we address the problem of automatically discovering atomic actions in unsupervised manner from instructional videos, which are rarely annotated with atomic actions. We present an unsupervised approach to learn atomic actions of structured human tasks from a variety of instructional videos based on a sequential stochastic autoregressive model for temporal segmentation of videos. This learns to represent and discover the sequential relationship between different atomic actions of the task, and which provides automatic and unsupervised self-labeling.

17.Reveal of Vision Transformers Robustness against Adversarial Attacks ⬇️

Attention-based networks have achieved state-of-the-art performance in many computer vision tasks, such as image classification. Unlike Convolutional Neural Network (CNN), the major part of the vanilla Vision Transformer (ViT) is the attention block that brings the power of mimicking the global context of the input image. This power is data hunger and hence, the larger the training data the better the performance. To overcome this limitation, many ViT-based networks, or hybrid-ViT, have been proposed to include local context during the training. The robustness of ViTs and its variants against adversarial attacks has not been widely invested in the literature. Some robustness attributes were revealed in few previous works and hence, more insight robustness attributes are yet unrevealed. This work studies the robustness of ViT variants 1) against different $L_p$-based adversarial attacks in comparison with CNNs and 2) under Adversarial Examples (AEs) after applying preprocessing defense methods. To that end, we run a set of experiments on 1000 images from ImageNet-1k and then provide an analysis that reveals that vanilla ViT or hybrid-ViT are more robust than CNNs. For instance, we found that 1) Vanilla ViTs or hybrid-ViTs are more robust than CNNs under $L_0$, $L_1$, $L_2$, $L_\infty$-based, and Color Channel Perturbations (CCP) attacks. 2) Vanilla ViTs are not responding to preprocessing defenses that mainly reduce the high frequency components while, hybrid-ViTs are more responsive to such defense. 3) CCP can be used as a preprocessing defense and larger ViT variants are found to be more responsive than other models. Furthermore, feature maps, attention maps, and Grad-CAM visualization jointly with image quality measures, and perturbations' energy spectrum are provided for an insight understanding of attention-based models.

18.Person Re-Identification with a Locally Aware Transformer ⬇️

Person Re-Identification is an important problem in computer vision-based surveillance applications, in which the same person is attempted to be identified from surveillance photographs in a variety of nearby zones. At present, the majority of Person re-ID techniques are based on Convolutional Neural Networks (CNNs), but Vision Transformers are beginning to displace pure CNNs for a variety of object recognition tasks. The primary output of a vision transformer is a global classification token, but vision transformers also yield local tokens which contain additional information about local regions of the image. Techniques to make use of these local tokens to improve classification accuracy are an active area of research. We propose a novel Locally Aware Transformer (LA-Transformer) that employs a Parts-based Convolution Baseline (PCB)-inspired strategy for aggregating globally enhanced local classification tokens into an ensemble of $\sqrt{N}$ classifiers, where $N$ is the number of patches. An additional novelty is that we incorporate blockwise fine-tuning which further improves re-ID accuracy. LA-Transformer with blockwise fine-tuning achieves rank-1 accuracy of $98.27 %$ with standard deviation of $0.13$ on the Market-1501 and $98.7%$ with standard deviation of $0.2$ on the CUHK03 dataset respectively, outperforming all other state-of-the-art published methods at the time of writing.

19.Incremental False Negative Detection for Contrastive Learning ⬇️

Self-supervised learning has recently shown great potential in vision tasks via contrastive learning, which aims to discriminate each image, or instance, in the dataset. However, such instance-level learning ignores the semantic relationship between instances and repels the anchor equally from the semantically similar samples, termed as false negatives. In this work, we first empirically highlight that the unfavorable effect from false negatives is more significant for the datasets containing images with more semantic concepts. To address the issue, we introduce a novel incremental false negative detection for self-supervised contrastive learning. Following the training process, when the encoder is gradually better-trained and the embedding space becomes more semantically structural, our method incrementally detects more reliable false negatives. Subsequently, during contrastive learning, we discuss two strategies to explicitly remove the detected false negatives. Extensive experiments show that our proposed method outperforms other self-supervised contrastive learning frameworks on multiple benchmarks within a limited compute.

20.Refiner: Refining Self-attention for Vision Transformers ⬇️

Vision Transformers (ViTs) have shown competitive accuracy in image classification tasks compared with CNNs. Yet, they generally require much more data for model pre-training. Most of recent works thus are dedicated to designing more complex architectures or training methods to address the data-efficiency issue of ViTs. However, few of them explore improving the self-attention mechanism, a key factor distinguishing ViTs from CNNs. Different from existing works, we introduce a conceptually simple scheme, called refiner, to directly refine the self-attention maps of ViTs. Specifically, refiner explores attention expansion that projects the multi-head attention maps to a higher-dimensional space to promote their diversity. Further, refiner applies convolutions to augment local patterns of the attention maps, which we show is equivalent to a distributed local attention features are aggregated locally with learnable kernels and then globally aggregated with self-attention. Extensive experiments demonstrate that refiner works surprisingly well. Significantly, it enables ViTs to achieve 86% top-1 classification accuracy on ImageNet with only 81M parameters.

21.Deep Learning 3D Dose Prediction for Conventional Lung IMRT Using Consistent/Unbiased Automated Plans ⬇️

Deep learning (DL) 3D dose prediction has recently gained a lot of attention. However, the variability of plan quality in the training dataset, generated manually by planners with wide range of expertise, can dramatically effect the quality of the final predictions. Moreover, any changes in the clinical criteria requires a new set of manually generated plans by planners to build a new prediction model. In this work, we instead use consistent plans generated by our in-house automated planning system (named ``ECHO'') to train the DL model. ECHO (expedited constrained hierarchical optimization) generates consistent/unbiased plans by solving large-scale constrained optimization problems sequentially. If the clinical criteria changes, a new training data set can be easily generated offline using ECHO, with no or limited human intervention, making the DL-based prediction model easily adaptable to the changes in the clinical practice. We used 120 conventional lung patients (100 for training, 20 for testing) with different beam configurations and trained our DL-model using manually-generated as well as automated ECHO plans. We evaluated different inputs: (1) CT+(PTV/OAR)contours, and (2) CT+contours+beam configurations, and different loss functions: (1) MAE (mean absolute error), and (2) MAE+DVH (dose volume histograms). The quality of the predictions was compared using different DVH metrics as well as dose-score and DVH-score, recently introduced by the AAPM knowledge-based planning grand challenge. The best results were obtained using automated ECHO plans and CT+contours+beam as training inputs and MAE+DVH as loss function.

22.Detection of marine floating plastic using Sentinel-2 imagery and machine learning models ⬇️

The increasing level of marine plastic pollution poses severe threats to the marine ecosystem and biodiversity. The present study attempted to explore the full functionality of open Sentinel satellite data and ML models for detecting and classifying floating plastic debris in Mytilene (Greece), Limassol (Cyprus), Calabria (Italy), and Beirut (Lebanon). Two ML models, i.e. Support Vector Machine (SVM) and Random Forest (RF) were utilized to carry out the classification analysis. In-situ plastic location data was collected from the control experiment conducted in Mytilene, Greece and Limassol, Cyprus, and the same was considered for training the models. Both remote sensing bands and spectral indices were used for developing the ML models. A spectral signature profile for plastic was created for discriminating the floating plastic from other marine debris. A newly developed index, kernel Normalized Difference Vegetation Index (kNDVI), was incorporated into the modelling to examine its contribution to model performances. Both SVM and RF were performed well in five models and test case combinations. Among the two ML models, the highest performance was measured for the RF. The inclusion of kNDVI was found effective and increased the model performances, reflected by high balanced accuracy measured for model 2 (~80% to ~98 % for SVM and ~87% to ~97 % for RF). Using the best-performed model, an automated floating plastic detection system was developed and tested in Calabria and Beirut. For both sites, the trained model had detected the floating plastic with ~99% accuracy. Among the six predictors, the FDI was found the most important variable for detecting marine floating plastic. These findings collectively suggest that high-resolution remote sensing imagery and the automated ML models can be an effective alternative for the cost-effective detection of marine floating plastic.

23.Open source disease analysis system of cactus by artificial intelligence and image processing ⬇️

There is a growing interest in cactus cultivation because of numerous cacti uses from houseplants to food and medicinal applications. Various diseases impact the growth of cacti. To develop an automated model for the analysis of cactus disease and to be able to quickly treat and prevent damage to the cactus. The Faster R-CNN and YOLO algorithm technique were used to analyze cactus diseases automatically distributed into six groups: 1) anthracnose, 2) canker, 3) lack of care, 4) aphid, 5) rusts and 6) normal group. Based on the experimental results the YOLOv5 algorithm was found to be more effective at detecting and identifying cactus disease than the Faster R-CNN algorithm. Data training and testing with YOLOv5S model resulted in a precision of 89.7% and an accuracy (recall) of 98.5%, which is effective enough for further use in a number of applications in cactus cultivation. Overall the YOLOv5 algorithm had a test time per image of only 26 milliseconds. Therefore, the YOLOv5 algorithm was found to suitable for mobile applications and this model could be further developed into a program for analyzing cactus disease.

24.Recovery Analysis for Plug-and-Play Priors using the Restricted Eigenvalue Condition ⬇️

The plug-and-play priors (PnP) and regularization by denoising (RED) methods have become widely used for solving inverse problems by leveraging pre-trained deep denoisers as image priors. While the empirical imaging performance and the theoretical convergence properties of these algorithms have been widely investigated, their recovery properties have not previously been theoretically analyzed. We address this gap by showing how to establish theoretical recovery guarantees for PnP/RED by assuming that the solution of these methods lies near the fixed-points of a deep neural network. We also present numerical results comparing the recovery performance of PnP/RED in compressive sensing against that of recent compressive sensing algorithms based on generative models. Our numerical results suggest that PnP with a pre-trained artifact removal network provides significantly better results compared to the existing state-of-the-art methods.

25.Shuffle Transformer: Rethinking Spatial Shuffle for Vision Transformer ⬇️

Very recently, Window-based Transformers, which computed self-attention within non-overlapping local windows, demonstrated promising results on image classification, semantic segmentation, and object detection. However, less study has been devoted to the cross-window connection which is the key element to improve the representation ability. In this work, we revisit the spatial shuffle as an efficient way to build connections among windows. As a result, we propose a new vision transformer, named Shuffle Transformer, which is highly efficient and easy to implement by modifying two lines of code. Furthermore, the depth-wise convolution is introduced to complement the spatial shuffle for enhancing neighbor-window connections. The proposed architectures achieve excellent performance on a wide range of visual tasks including image-level classification, object detection, and semantic segmentation. Code will be released for reproduction.

26.Efficient Iterative Amortized Inference for Learning Symmetric and Disentangled Multi-Object Representations ⬇️

Unsupervised multi-object representation learning depends on inductive biases to guide the discovery of object-centric representations that generalize. However, we observe that methods for learning these representations are either impractical due to long training times and large memory consumption or forego key inductive biases. In this work, we introduce EfficientMORL, an efficient framework for the unsupervised learning of object-centric representations. We show that optimization challenges caused by requiring both symmetry and disentanglement can in fact be addressed by high-cost iterative amortized inference by designing the framework to minimize its dependence on it. We take a two-stage approach to inference: first, a hierarchical variational autoencoder extracts symmetric and disentangled representations through bottom-up inference, and second, a lightweight network refines the representations with top-down feedback. The number of refinement steps taken during training is reduced following a curriculum, so that at test time with zero steps the model achieves 99.1% of the refined decomposition performance. We demonstrate strong object decomposition and disentanglement on the standard multi-object benchmark while achieving nearly an order of magnitude faster training and test time inference over the previous state-of-the-art model.

27.Adversarial Attack and Defense in Deep Ranking ⬇️

Deep Neural Network classifiers are vulnerable to adversarial attack, where an imperceptible perturbation could result in misclassification. However, the vulnerability of DNN-based image ranking systems remains under-explored. In this paper, we propose two attacks against deep ranking systems, i.e., Candidate Attack and Query Attack, that can raise or lower the rank of chosen candidates by adversarial perturbations. Specifically, the expected ranking order is first represented as a set of inequalities, and then a triplet-like objective function is designed to obtain the optimal perturbation. Conversely, an anti-collapse triplet defense is proposed to improve the ranking model robustness against all proposed attacks, where the model learns to prevent the positive and negative samples being pulled close to each other by adversarial attack. To comprehensively measure the empirical adversarial robustness of a ranking model with our defense, we propose an empirical robustness score, which involves a set of representative attacks against ranking models. Our adversarial ranking attacks and defenses are evaluated on MNIST, Fashion-MNIST, CUB200-2011, CARS196 and Stanford Online Products datasets. Experimental results demonstrate that a typical deep ranking system can be effectively compromised by our attacks. Nevertheless, our defense can significantly improve the ranking system robustness, and simultaneously mitigate a wide range of attacks.

28.End-to-end reconstruction meets data-driven regularization for inverse problems ⬇️

We propose an unsupervised approach for learning end-to-end reconstruction operators for ill-posed inverse problems. The proposed method combines the classical variational framework with iterative unrolling, which essentially seeks to minimize a weighted combination of the expected distortion in the measurement space and the Wasserstein-1 distance between the distributions of the reconstruction and ground-truth. More specifically, the regularizer in the variational setting is parametrized by a deep neural network and learned simultaneously with the unrolled reconstruction operator. The variational problem is then initialized with the reconstruction of the unrolled operator and solved iteratively till convergence. Notably, it takes significantly fewer iterations to converge, thanks to the excellent initialization obtained via the unrolled operator. The resulting approach combines the computational efficiency of end-to-end unrolled reconstruction with the well-posedness and noise-stability guarantees of the variational setting. Moreover, we demonstrate with the example of X-ray computed tomography (CT) that our approach outperforms state-of-the-art unsupervised methods, and that it outperforms or is on par with state-of-the-art supervised learned reconstruction approaches.

29.Multi-Exit Semantic Segmentation Networks ⬇️

Semantic segmentation arises as the backbone of many vision systems, spanning from self-driving cars and robot navigation to augmented reality and teleconferencing. Frequently operating under stringent latency constraints within a limited resource envelope, optimising for efficient execution becomes important. To this end, we propose a framework for converting state-of-the-art segmentation models to MESS networks; specially trained CNNs that employ parametrised early exits along their depth to save computation during inference on easier samples. Designing and training such networks naively can hurt performance. Thus, we propose a two-staged training process that pushes semantically important features early in the network. We co-optimise the number, placement and architecture of the attached segmentation heads, along with the exit policy, to adapt to the device capabilities and application-specific requirements. Optimising for speed, MESS networks can achieve latency gains of up to 2.83x over state-of-the-art methods with no accuracy degradation. Accordingly, optimising for accuracy, we achieve an improvement of up to 5.33 pp, under the same computational budget.

30.Self-supervised Depth Estimation Leveraging Global Perception and Geometric Smoothness Using On-board Videos ⬇️

Self-supervised depth estimation has drawn much attention in recent years as it does not require labeled data but image sequences. Moreover, it can be conveniently used in various applications, such as autonomous driving, robotics, realistic navigation, and smart cities. However, extracting global contextual information from images and predicting a geometrically natural depth map remain challenging. In this paper, we present DLNet for pixel-wise depth estimation, which simultaneously extracts global and local features with the aid of our depth Linformer block. This block consists of the Linformer and innovative soft split multi-layer perceptron blocks. Moreover, a three-dimensional geometry smoothness loss is proposed to predict a geometrically natural depth map by imposing the second-order smoothness constraint on the predicted three-dimensional point clouds, thereby realizing improved performance as a byproduct. Finally, we explore the multi-scale prediction strategy and propose the maximum margin dual-scale prediction strategy for further performance improvement. In experiments on the KITTI and Make3D benchmarks, the proposed DLNet achieves performance competitive to those of the state-of-the-art methods, reducing time and space complexities by more than $62%$ and $56%$, respectively. Extensive testing on various real-world situations further demonstrates the strong practicality and generalization capability of the proposed model.

31.The Distance Transform and its Computation ⬇️

Distance transformation is an image processing technique used for many different applications. Related to a binary image, the general idea is to determine the distance of all background points to the nearest object point (or vice versa). In this tutorial, different approaches are explained in detail and compared using examples. Corresponding source code is provided to facilitate own investigations. A particular objective of this tutorial is to clarify the difference between arbitrary distance transforms and exact Euclidean distance transformations.

32.Efficient training for future video generation based on hierarchical disentangled representation of latent variables ⬇️

Generating videos predicting the future of a given sequence has been an area of active research in recent years. However, an essential problem remains unsolved: most of the methods require large computational cost and memory usage for training. In this paper, we propose a novel method for generating future prediction videos with less memory usage than the conventional methods. This is a critical stepping stone in the path towards generating videos with high image quality, similar to that of generated images in the latest works in the field of image generation. We achieve high-efficiency by training our method in two stages: (1) image reconstruction to encode video frames into latent variables, and (2) latent variable prediction to generate the future sequence. Our method achieves an efficient compression of video into low-dimensional latent variables by decomposing each frame according to its hierarchical structure. That is, we consider that video can be separated into background and foreground objects, and that each object holds time-varying and time-independent information independently. Our experiments show that the proposed method can efficiently generate future prediction videos, even for complex datasets that cannot be handled by previous methods.

33.Self-Supervision & Meta-Learning for One-Shot Unsupervised Cross-Domain Detection ⬇️

Deep detection models have largely demonstrated to be extremely powerful in controlled settings, but appear brittle and fail when applied off-the-shelf on unseen domains. All the adaptive approaches developed to amend this issue access a sizable amount of target samples at training time, a strategy not suitable when the target is unknown and its data are not available in advance. Consider for instance the task of monitoring image feeds from social media: as every image is uploaded by a different user it belongs to a different target domain that is impossible to foresee during training. Our work addresses this setting, presenting an object detection algorithm able to perform unsupervised adaptation across domains by using only one target sample, seen at test time. We introduce a multi-task architecture that one-shot adapts to any incoming sample by iteratively solving a self-supervised task on it. We further exploit meta-learning to simulate single-sample cross domain learning episodes and better align to the test condition. Moreover, a cross-task pseudo-labeling procedure allows to focus on the image foreground and enhances the adaptation process. A thorough benchmark analysis against the most recent cross-domain detection methods and a detailed ablation study show the advantage of our approach.

34.Exploiting Emotional Dependencies with Graph Convolutional Networks for Facial Expression Recognition ⬇️

Over the past few years, deep learning methods have shown remarkable results in many face-related tasks including automatic facial expression recognition (FER) in-the-wild. Meanwhile, numerous models describing the human emotional states have been proposed by the psychology community. However, we have no clear evidence as to which representation is more appropriate and the majority of FER systems use either the categorical or the dimensional model of affect. Inspired by recent work in multi-label classification, this paper proposes a novel multi-task learning (MTL) framework that exploits the dependencies between these two models using a Graph Convolutional Network (GCN) to recognize facial expressions in-the-wild. Specifically, a shared feature representation is learned for both discrete and continuous recognition in a MTL setting. Moreover, the facial expression classifiers and the valence-arousal regressors are learned through a GCN that explicitly captures the dependencies between them. To evaluate the performance of our method under real-world conditions we train our models on AffectNet dataset. The results of our experiments show that our method outperforms the current state-of-the-art methods on discrete FER.

35.FINet: Dual Branches Feature Interaction for Partial-to-Partial Point Cloud Registration ⬇️

Data association is important in the point cloud registration. In this work, we propose to solve the partial-to-partial registration from a new perspective, by introducing feature interactions between the source and the reference clouds at the feature extraction stage, such that the registration can be realized without the explicit mask estimation or attentions for the overlapping detection as adopted previously. Specifically, we present FINet, a feature interaction-based structure with the capability to enable and strengthen the information associating between the inputs at multiple stages. To achieve this, we first split the features into two components, one for the rotation and one for the translation, based on the fact that they belong to different solution spaces, yielding a dual branches structure. Second, we insert several interaction modules at the feature extractor for the data association. Third, we propose a transformation sensitivity loss to obtain rotation-attentive and translation-attentive features. Experiments demonstrate that our method performs higher precision and robustness compared to the state-of-the-art traditional and learning-based methods.

36.Shape As Points: A Differentiable Poisson Solver ⬇️

In recent years, neural implicit representations gained popularity in 3D reconstruction due to their expressiveness and flexibility. However, the implicit nature of neural implicit representations results in slow inference time and requires careful initialization. In this paper, we revisit the classic yet ubiquitous point cloud representation and introduce a differentiable point-to-mesh layer using a differentiable formulation of Poisson Surface Reconstruction (PSR) that allows for a GPU-accelerated fast solution of the indicator function given an oriented point cloud. The differentiable PSR layer allows us to efficiently and differentiably bridge the explicit 3D point representation with the 3D mesh via the implicit indicator field, enabling end-to-end optimization of surface reconstruction metrics such as Chamfer distance. This duality between points and meshes hence allows us to represent shapes as oriented point clouds, which are explicit, lightweight and expressive. Compared to neural implicit representations, our Shape-As-Points (SAP) model is more interpretable, lightweight, and accelerates inference time by one order of magnitude. Compared to other explicit representations such as points, patches, and meshes, SAP produces topology-agnostic, watertight manifold surfaces. We demonstrate the effectiveness of SAP on the task of surface reconstruction from unoriented point clouds and learning-based reconstruction.

37.supervised adptive threshold network for instance segmentation ⬇️

Currently, instance segmentation is attracting more and more attention in machine learning region. However, there exists some defects on the information propagation in previous Mask R-CNN and other network models. In this paper, we propose supervised adaptive threshold network for instance segmentation. Specifically, we adopt the Mask R-CNN method based on adaptive threshold, and by establishing a layered adaptive network structure, it performs adaptive binarization on the probability graph generated by Mask RCNN to obtain better segmentation effect and reduce the error rate. At the same time, an adaptive feature pool is designed to make the transmission between different layers of the network more accurate and effective, reduce the loss in the process of feature transmission, and further improve the mask method. Experiments on benchmark data sets indicate that the effectiveness of the proposed model

38.Unsupervised Learning for Cuboid Shape Abstraction via Joint Segmentation from Point Clouds ⬇️

Representing complex 3D objects as simple geometric primitives, known as shape abstraction, is important for geometric modeling, structural analysis, and shape synthesis. In this paper, we propose an unsupervised shape abstraction method to map a point cloud into a compact cuboid representation. We jointly predict cuboid allocation as part segmentation and cuboid shapes and enforce the consistency between the segmentation and shape abstraction for self-learning. For the cuboid abstraction task, we transform the input point cloud into a set of parametric cuboids using a variational auto-encoder network. The segmentation network allocates each point into a cuboid considering the point-cuboid affinity. Without manual annotations of parts in point clouds, we design four novel losses to jointly supervise the two branches in terms of geometric similarity and cuboid compactness. We evaluate our method on multiple shape collections and demonstrate its superiority over existing shape abstraction methods. Moreover, based on our network architecture and learned representations, our approach supports various applications including structured shape generation, shape interpolation, and structural shape clustering.

39.Channel DropBlock: An Improved Regularization Method for Fine-Grained Visual Classification ⬇️

Classifying the sub-categories of an object from the same super-category (e.g., bird) in a fine-grained visual classification (FGVC) task highly relies on mining multiple discriminative features. Existing approaches mainly tackle this problem by introducing attention mechanisms to locate the discriminative parts or feature encoding approaches to extract the highly parameterized features in a weakly-supervised fashion. In this work, we propose a lightweight yet effective regularization method named Channel DropBlock (CDB), in combination with two alternative correlation metrics, to address this problem. The key idea is to randomly mask out a group of correlated channels during training to destruct features from co-adaptations and thus enhance feature representations. Extensive experiments on three benchmark FGVC datasets show that CDB effectively improves the performance.

40.Source-Free Open Compound Domain Adaptation in Semantic Segmentation ⬇️

In this work, we introduce a new concept, named source-free open compound domain adaptation (SF-OCDA), and study it in semantic segmentation. SF-OCDA is more challenging than the traditional domain adaptation but it is more practical. It jointly considers (1) the issues of data privacy and data storage and (2) the scenario of multiple target domains and unseen open domains. In SF-OCDA, only the source pre-trained model and the target data are available to learn the target model. The model is evaluated on the samples from the target and unseen open domains. To solve this problem, we present an effective framework by separating the training process into two stages: (1) pre-training a generalized source model and (2) adapting a target model with self-supervised learning. In our framework, we propose the Cross-Patch Style Swap (CPSS) to diversify samples with various patch styles in the feature-level, which can benefit the training of both stages. First, CPSS can significantly improve the generalization ability of the source model, providing more accurate pseudo-labels for the latter stage. Second, CPSS can reduce the influence of noisy pseudo-labels and also avoid the model overfitting to the target domain during self-supervised learning, consistently boosting the performance on the target and open domains. Experiments demonstrate that our method produces state-of-the-art results on the C-Driving dataset. Furthermore, our model also achieves the leading performance on CityScapes for domain generalization.

41.Multi-Target Domain Adaptation with Collaborative Consistency Learning ⬇️

Recently unsupervised domain adaptation for the semantic segmentation task has become more and more popular due to high-cost of pixel-level annotation on real-world images. However, most domain adaptation methods are only restricted to single-source-single-target pair, and can not be directly extended to multiple target domains. In this work, we propose a collaborative learning framework to achieve unsupervised multi-target domain adaptation. An unsupervised domain adaptation expert model is first trained for each source-target pair and is further encouraged to collaborate with each other through a bridge built between different target domains. These expert models are further improved by adding the regularization of making the consistent pixel-wise prediction for each sample with the same structured context. To obtain a single model that works across multiple target domains, we propose to simultaneously learn a student model which is trained to not only imitate the output of each expert on the corresponding target domain, but also to pull different expert close to each other with regularization on their weights. Extensive experiments demonstrate that the proposed method can effectively exploit rich structured information contained in both labeled source domain and multiple unlabeled target domains. Not only does it perform well across multiple target domains but also performs favorably against state-of-the-art unsupervised domain adaptation methods specially trained on a single source-target pair

42.Resolution learning in deep convolutional networks using scale-space theory ⬇️

Resolution in deep convolutional neural networks (CNNs) is typically bounded by the receptive field size through filter sizes, and subsampling layers or strided convolutions on feature maps. The optimal resolution may vary significantly depending on the dataset. Modern CNNs hard-code their resolution hyper-parameters in the network architecture which makes tuning such hyper-parameters cumbersome. We propose to do away with hard-coded resolution hyper-parameters and aim to learn the appropriate resolution from data. We use scale-space theory to obtain a self-similar parametrization of filters and make use of the N-Jet: a truncated Taylor series to approximate a filter by a learned combination of Gaussian derivative filters. The parameter {\sigma} of the Gaussian basis controls both the amount of detail the filter encodes and the spatial extent of the filter. Since {\sigma} is a continuous parameter, we can optimize it with respect to the loss. The proposed N-Jet layer achieves comparable performance when used in state-of-the art architectures, while learning the correct resolution in each layer automatically. We evaluate our N-Jet layer on both classification and segmentation, and we show that learning {\sigma} is especially beneficial for inputs at multiple sizes.

43.DINs: Deep Interactive Networks for Neurofibroma Segmentation in Neurofibromatosis Type 1 on Whole-Body MRI ⬇️

Neurofibromatosis type 1 (NF1) is an autosomal dominant tumor predisposition syndrome that involves the central and peripheral nervous systems. Accurate detection and segmentation of neurofibromas are essential for assessing tumor burden and longitudinal tumor size changes. Automatic convolutional neural networks (CNNs) are sensitive and vulnerable as tumors' variable anatomical location and heterogeneous appearance on MRI. In this study, we propose deep interactive networks (DINs) to address the above limitations. User interactions guide the model to recognize complicated tumors and quickly adapt to heterogeneous tumors. We introduce a simple but effective Exponential Distance Transform (ExpDT) that converts user interactions into guide maps regarded as the spatial and appearance prior. Comparing with popular Euclidean and geodesic distances, ExpDT is more robust to various image sizes, which reserves the distribution of interactive inputs. Furthermore, to enhance the tumor-related features, we design a deep interactive module to propagate the guides into deeper layers. We train and evaluate DINs on three MRI data sets from NF1 patients. The experiment results yield significant improvements of 44% and 14% in DSC comparing with automated and other interactive methods, respectively. We also experimentally demonstrate the efficiency of DINs in reducing user burden when comparing with conventional interactive methods. The source code of our method is available at \url{this https URL}.

44.ContourRender: Detecting Arbitrary Contour Shape For Instance Segmentation In One Pass ⬇️

Direct contour regression for instance segmentation is a challenging task. Previous works usually achieve it by learning to progressively refine the contour prediction or adopting a shape representation with limited expressiveness. In this work, we argue that the difficulty in regressing the contour points in one pass is mainly due to the ambiguity when discretizing a smooth contour into a polygon. To address the ambiguity, we propose a novel differentiable rendering-based approach named \textbf{ContourRender}. During training, it first predicts a contour generated by an invertible shape signature, and then optimizes the contour with the more stable silhouette by converting it to a contour mesh and rendering the mesh to a 2D map.
This method significantly improves the quality of contour without iterations or cascaded refinements. Moreover, as optimization is not needed during inference, the inference speed will not be influenced.
Experiments show the proposed ContourRender outperforms all the contour-based instance segmentation approaches on COCO, while stays competitive with the iteration-based state-of-the-art on Cityscapes. In addition, we specifically select a subset from COCO val2017 named COCO ContourHard-val to further demonstrate the contour quality improvements. Codes, models, and dataset split will be released.

45.ViTAE: Vision Transformer Advanced by Exploring Intrinsic Inductive Bias ⬇️

Transformers have shown great potential in various computer vision tasks owing to their strong capability in modeling long-range dependency using the self-attention mechanism. Nevertheless, vision transformers treat an image as 1D sequence of visual tokens, lacking an intrinsic inductive bias (IB) in modeling local visual structures and dealing with scale variance. Alternatively, they require large-scale training data and longer training schedules to learn the IB implicitly. In this paper, we propose a novel Vision Transformer Advanced by Exploring intrinsic IB from convolutions, \ie, ViTAE. Technically, ViTAE has several spatial pyramid reduction modules to downsample and embed the input image into tokens with rich multi-scale context by using multiple convolutions with different dilation rates. In this way, it acquires an intrinsic scale invariance IB and is able to learn robust feature representation for objects at various scales. Moreover, in each transformer layer, ViTAE has a convolution block in parallel to the multi-head self-attention module, whose features are fused and fed into the feed-forward network. Consequently, it has the intrinsic locality IB and is able to learn local features and global dependencies collaboratively. Experiments on ImageNet as well as downstream tasks prove the superiority of ViTAE over the baseline transformer and concurrent works. Source code and pretrained models will be available at GitHub.

46.Wide-Baseline Relative Camera Pose Estimation with Directional Learning ⬇️

Modern deep learning techniques that regress the relative camera pose between two images have difficulty dealing with challenging scenarios, such as large camera motions resulting in occlusions and significant changes in perspective that leave little overlap between images. These models continue to struggle even with the benefit of large supervised training datasets. To address the limitations of these models, we take inspiration from techniques that show regressing keypoint locations in 2D and 3D can be improved by estimating a discrete distribution over keypoint locations. Analogously, in this paper we explore improving camera pose regression by instead predicting a discrete distribution over camera poses. To realize this idea, we introduce DirectionNet, which estimates discrete distributions over the 5D relative pose space using a novel parameterization to make the estimation problem tractable. Specifically, DirectionNet factorizes relative camera pose, specified by a 3D rotation and a translation direction, into a set of 3D direction vectors. Since 3D directions can be identified with points on the sphere, DirectionNet estimates discrete distributions on the sphere as its output. We evaluate our model on challenging synthetic and real pose estimation datasets constructed from Matterport3D and InteriorNet. Promising results show a near 50% reduction in error over direct regression methods.

47.SelfDoc: Self-Supervised Document Representation Learning ⬇️

We propose SelfDoc, a task-agnostic pre-training framework for document image understanding. Because documents are multimodal and are intended for sequential reading, our framework exploits the positional, textual, and visual information of every semantically meaningful component in a document, and it models the contextualization between each block of content. Unlike existing document pre-training models, our model is coarse-grained instead of treating individual words as input, therefore avoiding an overly fine-grained with excessive contextualization. Beyond that, we introduce cross-modal learning in the model pre-training phase to fully leverage multimodal information from unlabeled documents. For downstream usage, we propose a novel modality-adaptive attention mechanism for multimodal feature fusion by adaptively emphasizing language and vision signals. Our framework benefits from self-supervised pre-training on documents without requiring annotations by a feature masking training strategy. It achieves superior performance on multiple downstream tasks with significantly fewer document images used in the pre-training stage compared to previous works.

48.Contextual Guided Segmentation Framework for Semi-supervised Video Instance Segmentation ⬇️

In this paper, we propose Contextual Guided Segmentation (CGS) framework for video instance segmentation in three passes. In the first pass, i.e., preview segmentation, we propose Instance Re-Identification Flow to estimate main properties of each instance (i.e., human/non-human, rigid/deformable, known/unknown category) by propagating its preview mask to other frames. In the second pass, i.e., contextual segmentation, we introduce multiple contextual segmentation schemes. For human instance, we develop skeleton-guided segmentation in a frame along with object flow to correct and refine the result across frames. For non-human instance, if the instance has a wide variation in appearance and belongs to known categories (which can be inferred from the initial mask), we adopt instance segmentation. If the non-human instance is nearly rigid, we train FCNs on synthesized images from the first frame of a video sequence. In the final pass, i.e., guided segmentation, we develop a novel fined-grained segmentation method on non-rectangular regions of interest (ROIs). The natural-shaped ROI is generated by applying guided attention from the neighbor frames of the current one to reduce the ambiguity in the segmentation of different overlapping instances. Forward mask propagation is followed by backward mask propagation to further restore missing instance fragments due to re-appeared instances, fast motion, occlusion, or heavy deformation. Finally, instances in each frame are merged based on their depth values, together with human and non-human object interaction and rare instance priority. Experiments conducted on the DAVIS Test-Challenge dataset demonstrate the effectiveness of our proposed framework. We achieved the 3rd consistently in the DAVIS Challenges 2017-2019 with 75.4%, 72.4%, and 78.4% in terms of global score, region similarity, and contour accuracy, respectively.

49.A Comprehensive Survey on Image Dehazing Based on Deep Learning ⬇️

The presence of haze significantly reduces the quality of images. Researchers have designed a variety of algorithms for image dehazing (ID) to restore the quality of hazy images. However, there are few studies that summarize the deep learning (DL) based dehazing technologies. In this paper, we conduct a comprehensive survey on the recent proposed dehazing methods. Firstly, we summarize the commonly used datasets, loss functions and evaluation metrics. Secondly, we group the existing researches of ID into two major categories: supervised ID and unsupervised ID. The core ideas of various influential dehazing models are introduced. Finally, the open issues for future research on ID are pointed out.

50.Exploring to establish an appropriate model for mage aesthetic assessment via CNN-based RSRL: An empirical study ⬇️

To establish an appropriate model for photo aesthetic assessment, in this paper, a D-measure which reflects the disentanglement degree of the final layer FC nodes of CNN is introduced. By combining F-measure with D-measure to obtain a FD measure, an algorithm of determining the optimal model from the multiple photo score prediction models generated by CNN-based repetitively self-revised learning(RSRL) is proposed. Furthermore, the first fixation perspective(FFP) and the assessment interest region(AIR) of the models are defined and calculated. The experimental results show that the FD measure is effective for establishing the appropriate model from the multiple score prediction models with different CNN structures. Moreover, the FD-determined optimal models with the comparatively high FD always have the FFP an AIR which are close to the human's aesthetic perception when enjoying photos.

51.Video Instance Segmentation using Inter-Frame Communication Transformers ⬇️

We propose a novel end-to-end solution for video instance segmentation (VIS) based on transformers. Recently, the per-clip pipeline shows superior performance over per-frame methods leveraging richer information from multiple frames. However, previous per-clip models require heavy computation and memory usage to achieve frame-to-frame communications, limiting practicality. In this work, we propose Inter-frame Communication Transformers (IFC), which significantly reduces the overhead for information-passing between frames by efficiently encoding the context within the input clip. Specifically, we propose to utilize concise memory tokens as a mean of conveying information as well as summarizing each frame scene. The features of each frame are enriched and correlated with other frames through exchange of information between the precisely encoded memory tokens. We validate our method on the latest benchmark sets and achieved the state-of-the-art performance (AP 44.6 on YouTube-VIS 2019 val set using the offline inference) while having a considerably fast runtime (89.4 FPS). Our method can also be applied to near-online inference for processing a video in real-time with only a small delay. The code will be made available.

52.Video Imprint ⬇️

A new unified video analytics framework (ER3) is proposed for complex event retrieval, recognition and recounting, based on the proposed video imprint representation, which exploits temporal correlations among image features across video frames. With the video imprint representation, it is convenient to reverse map back to both temporal and spatial locations in video frames, allowing for both key frame identification and key areas localization within each frame. In the proposed framework, a dedicated feature alignment module is incorporated for redundancy removal across frames to produce the tensor representation, i.e., the video imprint. Subsequently, the video imprint is individually fed into both a reasoning network and a feature aggregation module, for event recognition/recounting and event retrieval tasks, respectively. Thanks to its attention mechanism inspired by the memory networks used in language modeling, the proposed reasoning network is capable of simultaneous event category recognition and localization of the key pieces of evidence for event recounting. In addition, the latent structure in our reasoning network highlights the areas of the video imprint, which can be directly used for event recounting. With the event retrieval task, the compact video representation aggregated from the video imprint contributes to better retrieval results than existing state-of-the-art methods.

53.Highlighting the Importance of Reducing Research Bias and Carbon Emissions in CNNs ⬇️

Convolutional neural networks (CNNs) have become commonplace in addressing major challenges in computer vision. Researchers are not only coming up with new CNN architectures but are also researching different techniques to improve the performance of existing architectures. However, there is a tendency to over-emphasize performance improvement while neglecting certain important variables such as simplicity, versatility, the fairness of comparisons, and energy efficiency. Overlooking these variables in architectural design and evaluation has led to research bias and a significantly negative environmental impact. Furthermore, this can undermine the positive impact of research in using deep learning models to tackle climate change. Here, we perform an extensive and fair empirical study of a number of proposed techniques to gauge the utility of each technique for segmentation and classification. Our findings restate the importance of favoring simplicity over complexity in model design (Occam's Razor). Furthermore, our results indicate that simple standardized practices can lead to a significant reduction in environmental impact with little drop in performance. We highlight that there is a need to rethink the design and evaluation of CNNs to alleviate the issue of research bias and carbon emissions.

54.Few-shot segmentation of medical images based on meta-learning with implicit gradients ⬇️

Classical supervised methods commonly used often suffer from the requirement of an abudant number of training samples and are unable to generalize on unseen datasets. As a result, the broader application of any trained model is very limited in clinical settings. However, few-shot approaches can minimize the need for enormous reliable ground truth labels that are both labor intensive and expensive. To this end, we propose to exploit an optimization-based implicit model agnostic meta-learning {iMAML} algorithm in a few-shot setting for medical image segmentation. Our approach can leverage the learned weights from a diverse set of training samples and can be deployed on a new unseen dataset. We show that unlike classical few-shot learning approaches, our method has improved generalization capability. To our knowledge, this is the first work that exploits iMAML for medical image segmentation. Our quantitative results on publicly available skin and polyp datasets show that the proposed method outperforms the naive supervised baseline model and two recent few-shot segmentation approaches by large margins.

55.Alpha Matte Generation from Single Input for Portrait Matting ⬇️

Portrait matting is an important research problem with a wide range of applications, such as video conference app, image/video editing, and post-production. The goal is to predict an alpha matte that identifies the effect of each pixel on the foreground subject. Traditional approaches and most of the existing works utilized an additional input, e.g., trimap, background image, to predict alpha matte. However, providing additional input is not always practical. Besides, models are too sensitive to these additional inputs. In this paper, we introduce an additional input-free approach to perform portrait matting using Generative Adversarial Nets (GANs). We divide the main task into two subtasks. For this, we propose a segmentation network for the person segmentation and the alpha generation network for alpha matte prediction. While the segmentation network takes an input image and produces a coarse segmentation map, the alpha generation network utilizes the same input image as well as a coarse segmentation map that is produced by the segmentation network to predict the alpha matte. Besides, we present a segmentation encoding block to downsample the coarse segmentation map and provide feature representation to the residual block. Furthermore, we propose border loss to penalize only the borders of the subject separately which is more likely to be challenging and we also adapt perceptual loss for portrait matting. To train the proposed system, we combine two different popular training datasets to improve the amount of data as well as diversity to address domain shift problems in the inference time. We tested our model on three different benchmark datasets, namely Adobe Image Matting dataset, Portrait Matting dataset, and Distinctions dataset. The proposed method outperformed the MODNet method that also takes a single input.

56.Combinatorial Optimization for Panoptic Segmentation: An End-to-End Trainable Approach ⬇️

We propose an end-to-end trainable architecture for simultaneous semantic and instance segmentation (a.k.a. panoptic segmentation) consisting of a convolutional neural network and an asymmetric multiway cut problem solver. The latter solves a combinatorial optimization problem that elegantly incorporates semantic and boundary predictions to produce a panoptic labeling. Our formulation allows to directly maximize a smooth surrogate of the panoptic quality metric by backpropagating the gradient through the optimization problem. Experimental evaluation shows improvement of end-to-end learning w.r.t. comparable approaches on Cityscapes and COCO datasets. Overall, our approach shows the utility of using combinatorial optimization in tandem with deep learning in a challenging large scale real-world problem and showcases benefits and insights into training such an architecture end-to-end.

57.Transformer in Convolutional Neural Networks ⬇️

We tackle the low-efficiency flaw of vision transformer caused by the high computational/space complexity in Multi-Head Self-Attention (MHSA). To this end, we propose the Hierarchical MHSA (H-MHSA), whose representation is computed in a hierarchical manner. Specifically, our H-MHSA first learns feature relationships within small grids by viewing image patches as tokens. Then, small grids are merged into larger ones, within which feature relationship is learned by viewing each small grid at the preceding step as a token. This process is iterated to gradually reduce the number of tokens. The H-MHSA module is readily pluggable into any CNN architectures and amenable to training via backpropagation. We call this new backbone TransCNN, and it essentially inherits the advantages of both transformer and CNN. Experiments demonstrate that TransCNN achieves state-of-the-art accuracy for image recognition. Code and pretrained models are available at this https URL. This technical report will keep updating by adding more experiments.

58.Feature-based Style Randomization for Domain Generalization ⬇️

As a recent noticeable topic, domain generalization (DG) aims to first learn a generic model on multiple source domains and then directly generalize to an arbitrary unseen target domain without any additional adaption. In previous DG models, by generating virtual data to supplement observed source domains, the data augmentation based methods have shown its effectiveness. To simulate the possible unseen domains, most of them enrich the diversity of original data via image-level style transformation. However, we argue that the potential styles are hard to be exhaustively illustrated and fully augmented due to the limited referred styles, leading the diversity could not be always guaranteed. Unlike image-level augmentation, we in this paper develop a simple yet effective feature-based style randomization module to achieve feature-level augmentation, which can produce random styles via integrating random noise into the original style. Compared with existing image-level augmentation, our feature-level augmentation favors a more goal-oriented and sample-diverse way. Furthermore, to sufficiently explore the efficacy of the proposed module, we design a novel progressive training strategy to enable all parameters of the network to be fully trained. Extensive experiments on three standard benchmark datasets, i.e., PACS, VLCS and Office-Home, highlight the superiority of our method compared to the state-of-the-art methods.

59.Transformed ROIs for Capturing Visual Transformations in Videos ⬇️

Modeling the visual changes that an action brings to a scene is critical for video understanding. Currently, CNNs process one local neighbourhood at a time, so contextual relationships over longer ranges, while still learnable, are indirect. We present TROI, a plug-and-play module for CNNs to reason between mid-level feature representations that are otherwise separated in space and time. The module relates localized visual entities such as hands and interacting objects and transforms their corresponding regions of interest directly in the feature maps of convolutional layers. With TROI, we achieve state-of-the-art action recognition results on the large-scale datasets Something-Something-V2 and Epic-Kitchens-100.

60.Learning Video Models from Text: Zero-Shot Anticipation for Procedural Actions ⬇️

Can we teach a robot to recognize and make predictions for activities that it has never seen before? We tackle this problem by learning models for video from text. This paper presents a hierarchical model that generalizes instructional knowledge from large-scale text-corpora and transfers the knowledge to video. Given a portion of an instructional video, our model recognizes and predicts coherent and plausible actions multiple steps into the future, all in rich natural language. To demonstrate the capabilities of our model, we introduce the \emph{Tasty Videos Dataset V2}, a collection of 4022 recipes for zero-shot learning, recognition and anticipation. Extensive experiments with various evaluation metrics demonstrate the potential of our method for generalization, given limited video data for training models.

61.Technical Report: Temporal Aggregate Representations ⬇️

This technical report extends our work presented in [9] with more experiments. In [9], we tackle long-term video understanding, which requires reasoning from current and past or future observations and raises several fundamental questions. How should temporal or sequential relationships be modelled? What temporal extent of information and context needs to be processed? At what temporal scale should they be derived? [9] addresses these questions with a flexible multi-granular temporal aggregation framework. In this report, we conduct further experiments with this framework on different tasks and a new dataset, EPIC-KITCHENS-100.

62.Large-scale Unsupervised Semantic Segmentation ⬇️

Powered by the ImageNet dataset, unsupervised learning on large-scale data has made significant advances for classification tasks. There are two major challenges to allow such an attractive learning modality for segmentation tasks: i) a large-scale benchmark for assessing algorithms is missing; ii) unsupervised shape representation learning is difficult. We propose a new problem of large-scale unsupervised semantic segmentation (LUSS) with a newly created benchmark dataset to track the research progress. Based on the ImageNet dataset, we propose the ImageNet-S dataset with 1.2 million training images and 40k high-quality semantic segmentation annotations for evaluation. Our benchmark has a high data diversity and a clear task objective. We also present a simple yet effective baseline method that works surprisingly well for LUSS. In addition, we benchmark related un/weakly supervised methods accordingly, identifying the challenges and possible directions of LUSS.

63.Oriented Object Detection with Transformer ⬇️

Object detection with Transformers (DETR) has achieved a competitive performance over traditional detectors, such as Faster R-CNN. However, the potential of DETR remains largely unexplored for the more challenging task of arbitrary-oriented object detection problem. We provide the first attempt and implement Oriented Object DEtection with TRansformer ($\bf O^2DETR$) based on an end-to-end network. The contributions of $\rm O^2DETR$ include: 1) we provide a new insight into oriented object detection, by applying Transformer to directly and efficiently localize objects without a tedious process of rotated anchors as in conventional detectors; 2) we design a simple but highly efficient encoder for Transformer by replacing the attention mechanism with depthwise separable convolution, which can significantly reduce the memory and computational cost of using multi-scale features in the original Transformer; 3) our $\rm O^2DETR$ can be another new benchmark in the field of oriented object detection, which achieves up to 3.85 mAP improvement over Faster R-CNN and RetinaNet. We simply fine-tune the head mounted on $\rm O^2DETR$ in a cascaded architecture and achieve a competitive performance over SOTA in the DOTA dataset.

64.3D Convolution Neural Network based Person Identification using Gait cycles ⬇️

Human identification plays a prominent role in terms of security. In modern times security is becoming the key term for an individual or a country, especially for countries which are facing internal or external threats. Gait analysis is interpreted as the systematic study of the locomotive in humans. It can be used to extract the exact walking features of individuals. Walking features depends on biological as well as the physical feature of the object; hence, it is unique to every individual. In this work, gait features are used to identify an individual. The steps involve object detection, background subtraction, silhouettes extraction, skeletonization, and training 3D Convolution Neural Network on these gait features. The model is trained and evaluated on the dataset acquired by CASIA B Gait, which consists of 15000 videos of 124 subjects walking pattern captured from 11 different angles carrying objects such as bag and coat. The proposed method focuses more on the lower body part to extract features such as the angle between knee and thighs, hip angle, angle of contact, and many other features. The experimental results are compared with amongst accuracies of silhouettes as datasets for training and skeletonized image as training data. The results show that extracting the information from skeletonized data yields improved accuracy.

65.Go with the Flows: Mixtures of Normalizing Flows for Point Cloud Generation and Reconstruction ⬇️

Recently normalizing flows (NFs) have demonstrated state-of-the-art performance on modeling 3D point clouds while allowing sampling with arbitrary resolution at inference time. However, these flow-based models still require long training times and large models for representing complicated geometries. This work enhances their representational power by applying mixtures of NFs to point clouds. We show that in this more general framework each component learns to specialize in a particular subregion of an object in a completely unsupervised fashion. By instantiating each mixture component with a comparatively small NF we generate point clouds with improved details compared to single-flow-based models while using fewer parameters and considerably reducing the inference runtime. We further demonstrate that by adding data augmentation, individual mixture components can learn to specialize in a semantically meaningful manner. We evaluate mixtures of NFs on generation, autoencoding and single-view reconstruction based on the ShapeNet dataset.

66.MOC-GAN: Mixing Objects and Captions to Generate Realistic Images ⬇️

Generating images with conditional descriptions gains increasing interests in recent years. However, existing conditional inputs are suffering from either unstructured forms (captions) or limited information and expensive labeling (scene graphs). For a targeted scene, the core items, objects, are usually definite while their interactions are flexible and hard to clearly define. Thus, we introduce a more rational setting, generating a realistic image from the objects and captions. Under this setting, objects explicitly define the critical roles in the targeted images and captions implicitly describe their rich attributes and connections. Correspondingly, a MOC-GAN is proposed to mix the inputs of two modalities to generate realistic images. It firstly infers the implicit relations between object pairs from the captions to build a hidden-state scene graph. So a multi-layer representation containing objects, relations and captions is constructed, where the scene graph provides the structures of the scene and the caption provides the image-level guidance. Then a cascaded attentive generative network is designed to coarse-to-fine generate phrase patch by paying attention to the most relevant words in the caption. In addition, a phrase-wise DAMSM is proposed to better supervise the fine-grained phrase-patch consistency. On COCO dataset, our method outperforms the state-of-the-art methods on both Inception Score and FID while maintaining high visual quality. Extensive experiments demonstrate the unique features of our proposed method.

67.Rethinking Training from Scratch for Object Detection ⬇️

The ImageNet pre-training initialization is the de-facto standard for object detection. He et al. found it is possible to train detector from scratch(random initialization) while needing a longer training schedule with proper normalization technique. In this paper, we explore to directly pre-training on target dataset for object detection. Under this situation, we discover that the widely adopted large resizing strategy e.g. resize image to (1333, 800) is important for fine-tuning but it's not necessary for pre-training. Specifically, we propose a new training pipeline for object detection that follows `pre-training and fine-tuning', utilizing low resolution images within target dataset to pre-training detector then load it to fine-tuning with high resolution images. With this strategy, we can use batch normalization(BN) with large bath size during pre-training, it's also memory efficient that we can apply it on machine with very limited GPU memory(11G). We call it direct detection pre-training, and also use direct pre-training for short. Experiment results show that direct pre-training accelerates the pre-training phase by more than 11x on COCO dataset while with even +1.8mAP compared to ImageNet pre-training. Besides, we found direct pre-training is also applicable to transformer based backbones e.g. Swin Transformer. Code will be available.

68.Uformer: A General U-Shaped Transformer for Image Restoration ⬇️

In this paper, we present Uformer, an effective and efficient Transformer-based architecture, in which we build a hierarchical encoder-decoder network using the Transformer block for image restoration. Uformer has two core designs to make it suitable for this task. The first key element is a local-enhanced window Transformer block, where we use non-overlapping window-based self-attention to reduce the computational requirement and employ the depth-wise convolution in the feed-forward network to further improve its potential for capturing local context. The second key element is that we explore three skip-connection schemes to effectively deliver information from the encoder to the decoder. Powered by these two designs, Uformer enjoys a high capability for capturing useful dependencies for image restoration. Extensive experiments on several image restoration tasks demonstrate the superiority of Uformer, including image denoising, deraining, deblurring and demoireing. We expect that our work will encourage further research to explore Transformer-based architectures for low-level vision tasks. The code and models will be available at this https URL.

69.Deep Matching Prior: Test-Time Optimization for Dense Correspondence ⬇️

Conventional techniques to establish dense correspondences across visually or semantically similar images focused on designing a task-specific matching prior, which is difficult to model. To overcome this, recent learning-based methods have attempted to learn a good matching prior within a model itself on large training data. The performance improvement was apparent, but the need for sufficient training data and intensive learning hinders their applicability. Moreover, using the fixed model at test time does not account for the fact that a pair of images may require their own prior, thus providing limited performance and poor generalization to unseen images. In this paper, we show that an image pair-specific prior can be captured by solely optimizing the untrained matching networks on an input pair of images. Tailored for such test-time optimization for dense correspondence, we present a residual matching network and a confidence-aware contrastive loss to guarantee a meaningful convergence. Experiments demonstrate that our framework, dubbed Deep Matching Prior (DMP), is competitive, or even outperforms, against the latest learning-based methods on several benchmarks for geometric matching and semantic matching, even though it requires neither large training data nor intensive learning. With the networks pre-trained, DMP attains state-of-the-art performance on all benchmarks.

70.Referring Transformer: A One-step Approach to Multi-task Visual Grounding ⬇️

As an important step towards visual reasoning, visual grounding (e.g., phrase localization, referring expression comprehension/segmentation) has been widely explored Previous approaches to referring expression comprehension (REC) or segmentation (RES) either suffer from limited performance, due to a two-stage setup, or require the designing of complex task-specific one-stage architectures. In this paper, we propose a simple one-stage multi-task framework for visual grounding tasks. Specifically, we leverage a transformer architecture, where two modalities are fused in a visual-lingual encoder. In the decoder, the model learns to generate contextualized lingual queries which are then decoded and used to directly regress the bounding box and produce a segmentation mask for the corresponding referred regions. With this simple but highly contextualized model, we outperform state-of-the-arts methods by a large margin on both REC and RES tasks. We also show that a simple pre-training schedule (on an external dataset) further improves the performance. Extensive experiments and ablations illustrate that our model benefits greatly from contextualized information and multi-task training.

71.Reducing the feature divergence of RGB and near-infrared images using Switchable Normalization ⬇️

Visual pattern recognition over agricultural areas is an important application of aerial image processing. In this paper, we consider the multi-modality nature of agricultural aerial images and show that naively combining different modalities together without taking the feature divergence into account can lead to sub-optimal results. Thus, we apply a Switchable Normalization block to our DeepLabV3 segmentation model to alleviate the feature divergence. Using the popular symmetric Kullback Leibler divergence measure, we show that our model can greatly reduce the divergence between RGB and near-infrared channels. Together with a hybrid loss function, our model achieves nearly 10% improvements in mean IoU over previously published baseline.

72.Neural Implicit 3D Shapes from Single Images with Spatial Patterns ⬇️

3D shape reconstruction from a single image has been a long-standing problem in computer vision. The problem is ill-posed and highly challenging due to the information loss and occlusion that occurred during the imagery capture. In contrast to previous methods that learn holistic shape priors, we propose a method to learn spatial pattern priors for inferring the invisible regions of the underlying shape, wherein each 3D sample in the implicit shape representation is associated with a set of points generated by hand-crafted 3D mappings, along with their local image features. The proposed spatial pattern is significantly more informative and has distinctive descriptions on both visible and occluded locations. Most importantly, the key to our work is the ubiquitousness of the spatial patterns across shapes, which enables reasoning invisible parts of the underlying objects and thus greatly mitigates the occlusion issue. We devise a neural network that integrates spatial pattern representations and demonstrate the superiority of the proposed method on widely used metrics.

73.Multi-Level Graph Encoding with Structural-Collaborative Relation Learning for Skeleton-Based Person Re-Identification ⬇️

Skeleton-based person re-identification (Re-ID) is an emerging open topic providing great value for safety-critical applications. Existing methods typically extract hand-crafted features or model skeleton dynamics from the trajectory of body joints, while they rarely explore valuable relation information contained in body structure or motion. To fully explore body relations, we construct graphs to model human skeletons from different levels, and for the first time propose a Multi-level Graph encoding approach with Structural-Collaborative Relation learning (MG-SCR) to encode discriminative graph features for person Re-ID. Specifically, considering that structurally-connected body components are highly correlated in a skeleton, we first propose a multi-head structural relation layer to learn different relations of neighbor body-component nodes in graphs, which helps aggregate key correlative features for effective node representations. Second, inspired by the fact that body-component collaboration in walking usually carries recognizable patterns, we propose a cross-level collaborative relation layer to infer collaboration between different level components, so as to capture more discriminative skeleton graph features. Finally, to enhance graph dynamics encoding, we propose a novel self-supervised sparse sequential prediction task for model pre-training, which facilitates encoding high-level graph semantics for person Re-ID. MG-SCR outperforms state-of-the-art skeleton-based methods, and it achieves superior performance to many multi-modal methods that utilize extra RGB or depth features. Our codes are available at this https URL.

74.Using GANs to Augment Data for Cloud Image Segmentation Task ⬇️

While cloud/sky image segmentation has extensive real-world applications, a large amount of labelled data is needed to train a highly accurate models to perform the task. Scarcity of such volumes of cloud/sky images with corresponding ground-truth binary maps makes it highly difficult to train such complex image segmentation models. In this paper, we demonstrate the effectiveness of using Generative Adversarial Networks (GANs) to generate data to augment the training set in order to increase the prediction accuracy of image segmentation model. We further present a way to estimate ground-truth binary maps for the GAN-generated images to facilitate their effective use as augmented images. Finally, we validate our work with different statistical techniques.

75.Occlusion-aware Unsupervised Learning of Depth from 4-D Light Fields ⬇️

Depth estimation is a fundamental issue in 4-D light field processing and analysis. Although recent supervised learning-based light field depth estimation methods have significantly improved the accuracy and efficiency of traditional optimization-based ones, these methods rely on the training over light field data with ground-truth depth maps which are challenging to obtain or even unavailable for real-world light field data. Besides, due to the inevitable gap (or domain difference) between real-world and synthetic data, they may suffer from serious performance degradation when generalizing the models trained with synthetic data to real-world data. By contrast, we propose an unsupervised learning-based method, which does not require ground-truth depth as supervision during training. Specifically, based on the basic knowledge of the unique geometry structure of light field data, we present an occlusion-aware strategy to improve the accuracy on occlusion areas, in which we explore the angular coherence among subsets of the light field views to estimate initial depth maps, and utilize a constrained unsupervised loss to learn their corresponding reliability for final depth prediction. Additionally, we adopt a multi-scale network with a weighted smoothness loss to handle the textureless areas. Experimental results on synthetic data show that our method can significantly shrink the performance gap between the previous unsupervised method and supervised ones, and produce depth maps with comparable accuracy to traditional methods with obviously reduced computational cost. Moreover, experiments on real-world datasets show that our method can avoid the domain shift problem presented in supervised methods, demonstrating the great potential of our method.

76.SADRNet: Self-Aligned Dual Face Regression Networks for Robust 3D Dense Face Alignment and Reconstruction ⬇️

Three-dimensional face dense alignment and reconstruction in the wild is a challenging problem as partial facial information is commonly missing in occluded and large pose face images. Large head pose variations also increase the solution space and make the modeling more difficult. Our key idea is to model occlusion and pose to decompose this challenging task into several relatively more manageable subtasks. To this end, we propose an end-to-end framework, termed as Self-aligned Dual face Regression Network (SADRNet), which predicts a pose-dependent face, a pose-independent face. They are combined by an occlusion-aware self-alignment to generate the final 3D face. Extensive experiments on two popular benchmarks, AFLW2000-3D and Florence, demonstrate that the proposed method achieves significant superior performance over existing state-of-the-art methods.

77.An Adaptive Framework for Learning Unsupervised Depth Completion ⬇️

We present a method to infer a dense depth map from a color image and associated sparse depth measurements. Our main contribution lies in the design of an annealing process for determining co-visibility (occlusions, disocclusions) and the degree of regularization to impose on the model. We show that regularization and co-visibility are related via the fitness (residual) of model to data and both can be unified into a single framework to improve the learning process. Our method is an adaptive weighting scheme that guides optimization by measuring the residual at each pixel location over each training step for (i) estimating a soft visibility mask and (ii) determining the amount of regularization. We demonstrate the effectiveness our method by applying it to several recent unsupervised depth completion methods and improving their performance on public benchmark datasets, without incurring additional trainable parameters or increase in inference time. Code available at: this https URL.

78.Learning Topology from Synthetic Data for Unsupervised Depth Completion ⬇️

We present a method for inferring dense depth maps from images and sparse depth measurements by leveraging synthetic data to learn the association of sparse point clouds with dense natural shapes, and using the image as evidence to validate the predicted depth map. Our learned prior for natural shapes uses only sparse depth as input, not images, so the method is not affected by the covariate shift when attempting to transfer learned models from synthetic data to real ones. This allows us to use abundant synthetic data with ground truth to learn the most difficult component of the reconstruction process, which is topology estimation, and use the image to refine the prediction based on photometric evidence. Our approach uses fewer parameters than previous methods, yet, achieves the state of the art on both indoor and outdoor benchmark datasets. Code available at: this https URL.

79.Self-Damaging Contrastive Learning ⬇️

The recent breakthrough achieved by contrastive learning accelerates the pace for deploying unsupervised training on real-world data applications. However, unlabeled data in reality is commonly imbalanced and shows a long-tail distribution, and it is unclear how robustly the latest contrastive learning methods could perform in the practical scenario. This paper proposes to explicitly tackle this challenge, via a principled framework called Self-Damaging Contrastive Learning (SDCLR), to automatically balance the representation learning without knowing the classes. Our main inspiration is drawn from the recent finding that deep models have difficult-to-memorize samples, and those may be exposed through network pruning. It is further natural to hypothesize that long-tail samples are also tougher for the model to learn well due to insufficient examples. Hence, the key innovation in SDCLR is to create a dynamic self-competitor model to contrast with the target model, which is a pruned version of the latter. During training, contrasting the two models will lead to adaptive online mining of the most easily forgotten samples for the current target model, and implicitly emphasize them more in the contrastive loss. Extensive experiments across multiple datasets and imbalance settings show that SDCLR significantly improves not only overall accuracies but also balancedness, in terms of linear evaluation on the full-shot and few-shot settings. Our code is available at: this https URL.

80.Visual Search Asymmetry: Deep Nets and Humans Share Similar Inherent Biases ⬇️

Visual search is a ubiquitous and often challenging daily task, exemplified by looking for the car keys at home or a friend in a crowd. An intriguing property of some classical search tasks is an asymmetry such that finding a target A among distractors B can be easier than finding B among A. To elucidate the mechanisms responsible for asymmetry in visual search, we propose a computational model that takes a target and a search image as inputs and produces a sequence of eye movements until the target is found. The model integrates eccentricity-dependent visual recognition with target-dependent top-down cues. We compared the model against human behavior in six paradigmatic search tasks that show asymmetry in humans. Without prior exposure to the stimuli or task-specific training, the model provides a plausible mechanism for search asymmetry. We hypothesized that the polarity of search asymmetry arises from experience with the natural environment. We tested this hypothesis by training the model on an augmented version of ImageNet where the biases of natural images were either removed or reversed. The polarity of search asymmetry disappeared or was altered depending on the training protocol. This study highlights how classical perceptual properties can emerge in neural network models, without the need for task-specific training, but rather as a consequence of the statistical properties of the developmental diet fed to the model. All source code and stimuli are publicly available this https URL

81.Spectral Temporal Graph Neural Network for Trajectory Prediction ⬇️

An effective understanding of the contextual environment and accurate motion forecasting of surrounding agents is crucial for the development of autonomous vehicles and social mobile robots. This task is challenging since the behavior of an autonomous agent is not only affected by its own intention, but also by the static environment and surrounding dynamically interacting agents. Previous works focused on utilizing the spatial and temporal information in time domain while not sufficiently taking advantage of the cues in frequency domain. To this end, we propose a Spectral Temporal Graph Neural Network (SpecTGNN), which can capture inter-agent correlations and temporal dependency simultaneously in frequency domain in addition to time domain. SpecTGNN operates on both an agent graph with dynamic state information and an environment graph with the features extracted from context images in two streams. The model integrates graph Fourier transform, spectral graph convolution and temporal gated convolution to encode history information and forecast future trajectories. Moreover, we incorporate a multi-head spatio-temporal attention mechanism to mitigate the effect of error propagation in a long time horizon. We demonstrate the performance of SpecTGNN on two public trajectory prediction benchmark datasets, which achieves state-of-the-art performance in terms of prediction accuracy.

82.Feature Flow Regularization: Improving Structured Sparsity in Deep Neural Networks ⬇️

Pruning is a model compression method that removes redundant parameters in deep neural networks (DNNs) while maintaining accuracy. Most available filter pruning methods require complex treatments such as iterative pruning, features statistics/ranking, or additional optimization designs in the training process. In this paper, we propose a simple and effective regularization strategy from a new perspective of evolution of features, which we call feature flow regularization (FFR), for improving structured sparsity and filter pruning in DNNs. Specifically, FFR imposes controls on the gradient and curvature of feature flow along the neural network, which implicitly increases the sparsity of the parameters. The principle behind FFR is that coherent and smooth evolution of features will lead to an efficient network that avoids redundant parameters. The high structured sparsity obtained from FFR enables us to prune filters effectively. Experiments with VGGNets, ResNets on CIFAR-10/100, and Tiny ImageNet datasets demonstrate that FFR can significantly improve both unstructured and structured sparsity. Our pruning results in terms of reduction of parameters and FLOPs are comparable to or even better than those of state-of-the-art pruning methods.

83.Dynamic Resolution Network ⬇️

Deep convolutional neural networks (CNNs) are often of sophisticated design with numerous convolutional layers and learnable parameters for the accuracy reason. To alleviate the expensive costs of deploying them on mobile devices, recent works have made huge efforts for excavating redundancy in pre-defined architectures. Nevertheless, the redundancy on the input resolution of modern CNNs has not been fully investigated, i.e., the resolution of input image is fixed. In this paper, we observe that the smallest resolution for accurately predicting the given image is different using the same neural network. To this end, we propose a novel dynamic-resolution network (DRNet) in which the resolution is determined dynamically based on each input sample. Thus, a resolution predictor with negligible computational costs is explored and optimized jointly with the desired network. In practice, the predictor learns the smallest resolution that can retain and even exceed the original recognition accuracy for each image. During the inference, each input image will be resized to its predicted resolution for minimizing the overall computation burden. We then conduct extensive experiments on several benchmark networks and datasets. The results show that our DRNet can be embedded in any off-the-shelf network architecture to obtain a considerable reduction in computational complexity. For instance, DRNet achieves similar performance with an about 34% computation reduction, while gains 1.4% accuracy increase with 10% computation reduction compared to the original ResNet-50 on ImageNet.

84.Category Contrast for Unsupervised Domain Adaptation in Visual Tasks ⬇️

Instance contrast for unsupervised representation learning has achieved great success in recent years. In this work, we explore the idea of instance contrastive learning in unsupervised domain adaptation (UDA) and propose a novel Category Contrast technique (CaCo) that introduces semantic priors on top of instance discrimination for visual UDA tasks. By considering instance contrastive learning as a dictionary look-up operation, we construct a semantics-aware dictionary with samples from both source and target domains where each target sample is assigned a (pseudo) category label based on the category priors of source samples. This allows category contrastive learning (between target queries and the category-level dictionary) for category-discriminative yet domain-invariant feature representations: samples of the same category (from either source or target domain) are pulled closer while those of different categories are pushed apart simultaneously. Extensive UDA experiments in multiple visual tasks ($e.g.$, segmentation, classification and detection) show that the simple implementation of CaCo achieves superior performance as compared with the highly-optimized state-of-the-art methods. Analytically and empirically, the experiments also demonstrate that CaCo is complementary to existing UDA methods and generalizable to other learning setups such as semi-supervised learning, unsupervised model adaptation, etc.

85.RDA: Robust Domain Adaptation via Fourier Adversarial Attacking ⬇️

Unsupervised domain adaptation (UDA) involves a supervised loss in a labeled source domain and an unsupervised loss in an unlabeled target domain, which often faces more severe overfitting (than classical supervised learning) as the supervised source loss has clear domain gap and the unsupervised target loss is often noisy due to the lack of annotations. This paper presents RDA, a robust domain adaptation technique that introduces adversarial attacking to mitigate overfitting in UDA. We achieve robust domain adaptation by a novel Fourier adversarial attacking (FAA) method that allows large magnitude of perturbation noises but has minimal modification of image semantics, the former is critical to the effectiveness of its generated adversarial samples due to the existence of 'domain gaps'. Specifically, FAA decomposes images into multiple frequency components (FCs) and generates adversarial samples by just perturbating certain FCs that capture little semantic information. With FAA-generated samples, the training can continue the 'random walk' and drift into an area with a flat loss landscape, leading to more robust domain adaptation. Extensive experiments over multiple domain adaptation tasks show that RDA can work with different computer vision tasks with superior performance.

86.An End-to-End Breast Tumour Classification Model Using Context-Based Patch Modelling- A BiLSTM Approach for Image Classification ⬇️

Researchers working on computational analysis of Whole Slide Images (WSIs) in histopathology have primarily resorted to patch-based modelling due to large resolution of each WSI. The large resolution makes WSIs infeasible to be fed directly into the machine learning models due to computational constraints. However, due to patch-based analysis, most of the current methods fail to exploit the underlying spatial relationship among the patches. In our work, we have tried to integrate this relationship along with feature-based correlation among the extracted patches from the particular tumorous region. For the given task of classification, we have used BiLSTMs to model both forward and backward contextual relationship. RNN based models eliminate the limitation of sequence size by allowing the modelling of variable size images within a deep learning model. We have also incorporated the effect of spatial continuity by exploring different scanning techniques used to sample patches. To establish the efficiency of our approach, we trained and tested our model on two datasets, microscopy images and WSI tumour regions. After comparing with contemporary literature we achieved the better performance with accuracy of 90% for microscopy image dataset. For WSI tumour region dataset, we compared the classification results with deep learning networks such as ResNet, DenseNet, and InceptionV3 using maximum voting technique. We achieved the highest performance accuracy of 84%. We found out that BiLSTMs with CNN features have performed much better in modelling patches into an end-to-end Image classification network. Additionally, the variable dimensions of WSI tumour regions were used for classification without the need for resizing. This suggests that our method is independent of tumour image size and can process large dimensional images without losing the resolution details.

87.Convolutional Neural Networks with Gated Recurrent Connections ⬇️

The convolutional neural network (CNN) has become a basic model for solving many computer vision problems. In recent years, a new class of CNNs, recurrent convolution neural network (RCNN), inspired by abundant recurrent connections in the visual systems of animals, was proposed. The critical element of RCNN is the recurrent convolutional layer (RCL), which incorporates recurrent connections between neurons in the standard convolutional layer. With increasing number of recurrent computations, the receptive fields (RFs) of neurons in RCL expand unboundedly, which is inconsistent with biological facts. We propose to modulate the RFs of neurons by introducing gates to the recurrent connections. The gates control the amount of context information inputting to the neurons and the neurons' RFs therefore become adaptive. The resulting layer is called gated recurrent convolution layer (GRCL). Multiple GRCLs constitute a deep model called gated RCNN (GRCNN). The GRCNN was evaluated on several computer vision tasks including object recognition, scene text recognition and object detection, and obtained much better results than the RCNN. In addition, when combined with other adaptive RF techniques, the GRCNN demonstrated competitive performance to the state-of-the-art models on benchmark datasets for these tasks. The codes are released at \href{this https URL}{this https URL}.

88.Region-aware Adaptive Instance Normalization for Image Harmonization ⬇️

Image composition plays a common but important role in photo editing. To acquire photo-realistic composite images, one must adjust the appearance and visual style of the foreground to be compatible with the background. Existing deep learning methods for harmonizing composite images directly learn an image mapping network from the composite to the real one, without explicit exploration on visual style consistency between the background and the foreground images. To ensure the visual style consistency between the foreground and the background, in this paper, we treat image harmonization as a style transfer problem. In particular, we propose a simple yet effective Region-aware Adaptive Instance Normalization (RAIN) module, which explicitly formulates the visual style from the background and adaptively applies them to the foreground. With our settings, our RAIN module can be used as a drop-in module for existing image harmonization networks and is able to bring significant improvements. Extensive experiments on the existing image harmonization benchmark datasets show the superior capability of the proposed method. Code is available at {this https URL}.

89.Patch Slimming for Efficient Vision Transformers ⬇️

This paper studies the efficiency problem for visual transformers by excavating redundant calculation in given networks. The recent transformer architecture has demonstrated its effectiveness for achieving excellent performance on a series of computer vision tasks. However, similar to that of convolutional neural networks, the huge computational cost of vision transformers is still a severe issue. Considering that the attention mechanism aggregates different patches layer-by-layer, we present a novel patch slimming approach that discards useless patches in a top-down paradigm. We first identify the effective patches in the last layer and then use them to guide the patch selection process of previous layers. For each layer, the impact of a patch on the final output feature is approximated and patches with less impact will be removed. Experimental results on benchmark datasets demonstrate that the proposed method can significantly reduce the computational costs of vision transformers without affecting their performances. For example, over 45% FLOPs of the ViT-Ti model can be reduced with only 0.2% top-1 accuracy drop on the ImageNet dataset.

90.Semi-Supervised Domain Adaptation via Adaptive and Progressive Feature Alignment ⬇️

Contemporary domain adaptive semantic segmentation aims to address data annotation challenges by assuming that target domains are completely unannotated. However, annotating a few target samples is usually very manageable and worthwhile especially if it improves the adaptation performance substantially. This paper presents SSDAS, a Semi-Supervised Domain Adaptive image Segmentation network that employs a few labeled target samples as anchors for adaptive and progressive feature alignment between labeled source samples and unlabeled target samples. We position the few labeled target samples as references that gauge the similarity between source and target features and guide adaptive inter-domain alignment for learning more similar source features. In addition, we replace the dissimilar source features by high-confidence target features continuously during the iterative training process, which achieves progressive intra-domain alignment between confident and unconfident target features. Extensive experiments show the proposed SSDAS greatly outperforms a number of baselines, i.e., UDA-based semantic segmentation and SSDA-based image classification. In addition, SSDAS is complementary and can be easily incorporated into UDA-based methods with consistent improvements in domain adaptive semantic segmentation.

91.Multi-Camera Vehicle Counting Using Edge-AI ⬇️

This paper presents a novel solution to automatically count vehicles in a parking lot using images captured by smart cameras. Unlike most of the literature on this task, which focuses on the analysis of single images, this paper proposes the use of multiple visual sources to monitor a wider parking area from different perspectives. The proposed multi-camera system is capable of automatically estimate the number of cars present in the entire parking lot directly on board the edge devices. It comprises an on-device deep learning-based detector that locates and counts the vehicles from the captured images and a decentralized geometric-based approach that can analyze the inter-camera shared areas and merge the data acquired by all the devices. We conduct the experimental evaluation on an extended version of the CNRPark-EXT dataset, a collection of images taken from the parking lot on the campus of the National Research Council (CNR) in Pisa, Italy. We show that our system is robust and takes advantage of the redundant information deriving from the different cameras, improving the overall performance without requiring any extra geometrical information of the monitored scene.

92.Making CNNs Interpretable by Building Dynamic Sequential Decision Forests with Top-down Hierarchy Learning ⬇️

In this paper, we propose a generic model transfer scheme to make Convlutional Neural Networks (CNNs) interpretable, while maintaining their high classification accuracy. We achieve this by building a differentiable decision forest on top of CNNs, which enjoys two characteristics: 1) During training, the tree hierarchies of the forest are learned in a top-down manner under the guidance from the category semantics embedded in the pre-trained CNN weights; 2) During inference, a single decision tree is dynamically selected from the forest for each input sample, enabling the transferred model to make sequential decisions corresponding to the attributes shared by semantically-similar categories, rather than directly performing flat classification. We name the transferred model deep Dynamic Sequential Decision Forest (dDSDF). Experimental results show that dDSDF not only achieves higher classification accuracy than its conuterpart, i.e., the original CNN, but has much better interpretability, as qualitatively it has plausible hierarchies and quantitatively it leads to more precise saliency maps.

93.Web based disease prediction and recommender system ⬇️

Worldwide, several cases go undiagnosed due to poor healthcare support in remote areas. In this context, a centralized system is needed for effective monitoring and analysis of the medical records. A web-based patient diagnostic system is a central platform to store the medical history and predict the possible disease based on the current symptoms experienced by a patient to ensure faster and accurate diagnosis. Early disease prediction can help the users determine the severity of the disease and take quick action. The proposed web-based disease prediction system utilizes machine learning based classification techniques on a data set acquired from the National Centre of Disease Control (NCDC). $K$-nearest neighbor (K-NN), random forest and naive bayes classification approaches are utilized and an ensemble voting algorithm is also proposed where each classifier is assigned weights dynamically based on the prediction confidence. The proposed system is also equipped with a recommendation scheme to recommend the type of tests based on the existing symptoms of the patient, so that necessary precautions can be taken. A centralized database ensures that the medical data is preserved and there is transparency in the system. The tampering into the system is prevented by giving the no "updation" rights once the diagnosis is created.

94.T-Net: Deep Stacked Scale-Iteration Network for Image Dehazing ⬇️

Hazy images reduce the visibility of the image content, and haze will lead to failure in handling subsequent computer vision tasks. In this paper, we address the problem of image dehazing by proposing a dehazing network named T-Net, which consists of a backbone network based on the U-Net architecture and a dual attention module. And it can achieve multi-scale feature fusion by using skip connections with a new fusion strategy. Furthermore, by repeatedly unfolding the plain T-Net, Stack T-Net is proposed to take advantage of the dependence of deep features across stages via a recursive strategy. In order to reduce network parameters, the intra-stage recursive computation of ResNet is adopted in our Stack T-Net. And we take both the stage-wise result and the original hazy image as input to each T-Net and finally output the prediction of clean image. Experimental results on both synthetic and real-world images demonstrate that our plain T-Net and the advanced Stack T-Net perform favorably against the state-of-the-art dehazing algorithms, and show that our Stack T-Net could further improve the dehazing effect, demonstrating the effectiveness of the recursive strategy.

95.Points2Polygons: Context-Based Segmentation from Weak Labels Using Adversarial Networks ⬇️

In applied image segmentation tasks, the ability to provide numerous and precise labels for training is paramount to the accuracy of the model at inference time. However, this overhead is often neglected, and recently proposed segmentation architectures rely heavily on the availability and fidelity of ground truth labels to achieve state-of-the-art accuracies. Failure to acknowledge the difficulty in creating adequate ground truths can lead to an over-reliance on pre-trained models or a lack of adoption in real-world applications. We introduce Points2Polygons (P2P), a model which makes use of contextual metric learning techniques that directly addresses this problem. Points2Polygons performs well against existing fully-supervised segmentation baselines with limited training data, despite using lightweight segmentation models (U-Net with a ResNet18 backbone) and having access to only weak labels in the form of object centroids and no pre-training. We demonstrate this on several different small but non-trivial datasets. We show that metric learning using contextual data provides key insights for self-supervised tasks in general, and allow segmentation models to easily generalize across traditionally label-intensive domains in computer vision.

96.IPS300+: a Challenging Multimodal Dataset for Intersection Perception System ⬇️

Due to the high complexity and occlusion, insufficient perception in the crowded urban intersection can be a serious safety risk for both human drivers and autonomous algorithms, whereas CVIS (Cooperative Vehicle Infrastructure System) is a proposed solution for full-participants perception in this scenario. However, the research on roadside multimodal perception is still in its infancy, and there is no open-source dataset for such scenario. Accordingly, this paper fills the gap. Through an IPS (Intersection Perception System) installed at the diagonal of the intersection, this paper proposes a high-quality multimodal dataset for the intersection perception task. The center of the experimental intersection covers an area of 3000m2, and the extended distance reaches 300m, which is typical for CVIS. The first batch of open-source data includes 14198 frames, and each frame has an average of 319.84 labels, which is 9.6 times larger than the most crowded dataset (H3D dataset in 2019) by now. In order to facilitate further study, this dataset tries to keep the label documents consistent with the KITTI dataset, and a standardized benchmark is created for algorithm evaluation. Our dataset is available at: this http URL.

97.Radar-Camera Pixel Depth Association for Depth Completion ⬇️

While radar and video data can be readily fused at the detection level, fusing them at the pixel level is potentially more beneficial. This is also more challenging in part due to the sparsity of radar, but also because automotive radar beams are much wider than a typical pixel combined with a large baseline between camera and radar, which results in poor association between radar pixels and color pixel. A consequence is that depth completion methods designed for LiDAR and video fare poorly for radar and video. Here we propose a radar-to-pixel association stage which learns a mapping from radar returns to pixels. This mapping also serves to densify radar returns. Using this as a first stage, followed by a more traditional depth completion method, we are able to achieve image-guided depth completion with radar and video. We demonstrate performance superior to camera and radar alone on the nuScenes dataset. Our source code is available at this https URL.

98.Visual communication of object concepts at different levels of abstraction ⬇️

People can produce drawings of specific entities (e.g., Garfield), as well as general categories (e.g., "cat"). What explains this ability to produce such varied drawings of even highly familiar object concepts? We hypothesized that drawing objects at different levels of abstraction depends on both sensory information and representational goals, such that drawings intended to portray a recently seen object preserve more detail than those intended to represent a category. Participants drew objects cued either with a photo or a category label. For each cue type, half the participants aimed to draw a specific exemplar; the other half aimed to draw the category. We found that label-cued category drawings were the most recognizable at the basic level, whereas photo-cued exemplar drawings were the least recognizable. Together, these findings highlight the importance of task context for explaining how people use drawings to communicate visual concepts in different ways.

99.GLSD: The Global Large-Scale Ship Database and Baseline Evaluations ⬇️

In this paper, we introduce a challenging global large-scale ship database (called GLSD), designed specifically for ship detection tasks. The designed GLSD database includes a total of 140,616 annotated instances from 100,729 images. Based on the collected images, we propose 13 categories that widely exists in international routes. These categories include sailing boat, fishing boat, passenger ship, war ship, general cargo ship, container ship, bulk cargo carrier, barge, ore carrier, speed boat, canoe, oil carrier, and tug. The motivations of developing GLSD include the following: 1) providing a refined ship detection database; 2) providing the worldwide researchers of ship detection and exhaustive label information (bounding box and ship class label) in one uniform global database; and 3) providing a large-scale ship database with geographic information (port and country information) that benefits multi-modal analysis. In addition, we discuss the evaluation protocols given image characteristics in GLSD and analyze the performance of selected state-of-the-art object detection algorithms on GSLD, providing baselines for future studies. More information regarding the designed GLSD can be found at this https URL.

100.Predify: Augmenting deep neural networks with brain-inspired predictive coding dynamics ⬇️

Deep neural networks excel at image classification, but their performance is far less robust to input perturbations than human perception. In this work we explore whether this shortcoming may be partly addressed by incorporating brain-inspired recurrent dynamics in deep convolutional networks. We take inspiration from a popular framework in neuroscience: 'predictive coding'. At each layer of the hierarchical model, generative feedback 'predicts' (i.e., reconstructs) the pattern of activity in the previous layer. The reconstruction errors are used to iteratively update the network's representations across timesteps, and to optimize the network's feedback weights over the natural image dataset-a form of unsupervised training. We show that implementing this strategy into two popular networks, VGG16 and EfficientNetB0, improves their robustness against various corruptions. We hypothesize that other feedforward networks could similarly benefit from the proposed framework. To promote research in this direction, we provide an open-sourced PyTorch-based package called Predify, which can be used to implement and investigate the impacts of the predictive coding dynamics in any convolutional neural network.

101.ZeroWaste Dataset: Towards Automated Waste Recycling ⬇️

Less than 35% of recyclable waste is being actually recycled in the US, which leads to increased soil and sea pollution and is one of the major concerns of environmental researchers as well as the common public. At the heart of the problem is the inefficiencies of the waste sorting process (separating paper, plastic, metal, glass, etc.) due to the extremely complex and cluttered nature of the waste stream. Automated waste detection strategies have a great potential to enable more efficient, reliable and safer waste sorting practices, but the literature lacks comprehensive datasets and methodology for the industrial waste sorting solutions. In this paper, we take a step towards computer-aided waste detection and present the first in-the-wild industrial-grade waste detection and segmentation dataset, ZeroWaste. This dataset contains over1800fully segmented video frames collected from a real waste sorting plant along with waste material labels for training and evaluation of the segmentation methods, as well as over6000unlabeled frames that can be further used for semi-supervised and self-supervised learning techniques. ZeroWaste also provides frames of the conveyor belt before and after the sorting process, comprising a novel setup that can be used for weakly-supervised segmentation. We present baselines for fully-, semi- and weakly-supervised segmentation methods. Our experimental results demonstrate that state-of-the-art segmentation methods struggle to correctly detect and classify target objects which suggests the challenging nature of our proposed in-the-wild dataset. We believe that ZeroWastewill catalyze research in object detection and semantic segmentation in extreme clutter as well as applications in the recycling domain. Our project page can be found athttp://ai.this http URL.

102.DISCO: accurate Discrete Scale Convolutions ⬇️

Scale is often seen as a given, disturbing factor in many vision tasks. When doing so it is one of the factors why we need more data during learning. In recent work scale equivariance was added to convolutional neural networks. It was shown to be effective for a range of tasks. We aim for accurate scale-equivariant convolutional neural networks (SE-CNNs) applicable for problems where high granularity of scale and small filter sizes are required. Current SE-CNNs rely on weight sharing and filter rescaling, the latter of which is accurate for integer scales only. To reach accurate scale equivariance, we derive general constraints under which scale-convolution remains equivariant to discrete rescaling. We find the exact solution for all cases where it exists, and compute the approximation for the rest. The discrete scale-convolution pays off, as demonstrated in a new state-of-the-art classification on MNIST-scale and improving the results on STL-10. With the same SE scheme, we also improve the computational effort of a scale-equivariant Siamese tracker on OTB-13.

103.Hierarchical Video Generation for Complex Data ⬇️

Videos can often be created by first outlining a global description of the scene and then adding local details. Inspired by this we propose a hierarchical model for video generation which follows a coarse to fine approach. First our model generates a low resolution video, establishing the global scene structure, that is then refined by subsequent levels in the hierarchy. We train each level in our hierarchy sequentially on partial views of the videos. This reduces the computational complexity of our generative model, which scales to high-resolution videos beyond a few frames. We validate our approach on Kinetics-600 and BDD100K, for which we train a three level model capable of generating 256x256 videos with 48 frames.

104.Hidden Markov Modeling for Maximum Likelihood Neuron Reconstruction ⬇️

Recent advances in brain clearing and imaging have made it possible to image entire mammalian brains at sub-micron resolution. These images offer the potential to assemble brain-wide atlases of projection neuron morphology, but manual neuron reconstruction remains a bottleneck. Here we present a method inspired by hidden Markov modeling and appearance modeling of fluorescent neuron images that can automatically trace neuronal processes. Our method leverages dynamic programming to scale to terabyte sized image data and can be applied to images with one or more neurons. We applied our algorithm to the output of image segmentation models where false negatives severed neuronal processes, and showed that it can follow axons in the presence of noise or nearby neurons. Our method has the potential to be integrated into a semi or fully automated reconstruction pipeline. Additionally, it creates a framework through which users can intervene with hard constraints to, for example, rule out certain reconstructions, or assign axons to particular cell bodies.

105.Efficient Classification of Very Large Images with Tiny Objects ⬇️

An increasing number of applications in the computer vision domain, specially, in medical imaging and remote sensing, are challenging when the goal is to classify very large images with tiny objects. More specifically, these type of classification tasks face two key challenges: $i$) the size of the input image in the target dataset is usually in the order of megapixels, however, existing deep architectures do not easily operate on such big images due to memory constraints, consequently, we seek a memory-efficient method to process these images; and $ii$) only a small fraction of the input images are informative of the label of interest, resulting in low region of interest (ROI) to image ratio. However, most of the current convolutional neural networks (CNNs) are designed for image classification datasets that have relatively large ROIs and small image size (sub-megapixel). Existing approaches have addressed these two challenges in isolation. We present an end-to-end CNN model termed Zoom-In network that leverages hierarchical attention sampling for classification of large images with tiny objects using a single GPU. We evaluate our method on two large-image datasets and one gigapixel dataset. Experimental results show that our model achieves higher accuracy than existing methods while requiring less computing resources.

106.RegionViT: Regional-to-Local Attention for Vision Transformers ⬇️

Vision transformer (ViT) has recently showed its strong capability in achieving comparable results to convolutional neural networks (CNNs) on image classification. However, vanilla ViT simply inherits the same architecture from the natural language processing directly, which is often not optimized for vision applications. Motivated by this, in this paper, we propose a new architecture that adopts the pyramid structure and employ a novel regional-to-local attention rather than global self-attention in vision transformers. More specifically, our model first generates regional tokens and local tokens from an image with different patch sizes, where each regional token is associated with a set of local tokens based on the spatial location. The regional-to-local attention includes two steps: first, the regional self-attention extract global information among all regional tokens and then the local self-attention exchanges the information among one regional token and the associated local tokens via self-attention. Therefore, even though local self-attention confines the scope in a local region but it can still receive global information. Extensive experiments on three vision tasks, including image classification, object detection and action recognition, show that our approach outperforms or is on par with state-of-the-art ViT variants including many concurrent works. Our source codes and models will be publicly available.

107.GAN Cocktail: mixing GANs without dataset access ⬇️

Today's generative models are capable of synthesizing high-fidelity images, but each model specializes on a specific target domain. This raises the need for model merging: combining two or more pretrained generative models into a single unified one. In this work we tackle the problem of model merging, given two constraints that often come up in the real world: (1) no access to the original training data, and (2) without increasing the size of the neural network. To the best of our knowledge, model merging under these constraints has not been studied thus far. We propose a novel, two-stage solution. In the first stage, we transform the weights of all the models to the same parameter space by a technique we term model rooting. In the second stage, we merge the rooted models by averaging their weights and fine-tuning them for each specific domain, using only data generated by the original trained models. We demonstrate that our approach is superior to baseline methods and to existing transfer learning techniques, and investigate several applications.

108.Active Speaker Detection as a Multi-Objective Optimization with Uncertainty-based Multimodal Fusion ⬇️

It is now well established from a variety of studies that there is a significant benefit from combining video and audio data in detecting active speakers. However, either of the modalities can potentially mislead audiovisual fusion by inducing unreliable or deceptive information. This paper outlines active speaker detection as a multi-objective learning problem to leverage best of each modalities using a novel self-attention, uncertainty-based multimodal fusion scheme. Results obtained show that the proposed multi-objective learning architecture outperforms traditional approaches in improving both mAP and AUC scores. We further demonstrate that our fusion strategy surpasses, in active speaker detection, other modality fusion methods reported in various disciplines. We finally show that the proposed method significantly improves the state-of-the-art on the AVA-ActiveSpeaker dataset.

109.Deep Medial Fields ⬇️

Implicit representations of geometry, such as occupancy fields or signed distance fields (SDF), have recently re-gained popularity in encoding 3D solid shape in a functional form. In this work, we introduce medial fields: a field function derived from the medial axis transform (MAT) that makes available information about the underlying 3D geometry that is immediately useful for a number of downstream tasks. In particular, the medial field encodes the local thickness of a 3D shape, and enables O(1) projection of a query point onto the medial axis. To construct the medial field we require nothing but the SDF of the shape itself, thus allowing its straightforward incorporation in any application that relies on signed distance fields. Working in unison with the O(1) surface projection supported by the SDF, the medial field opens the door for an entirely new set of efficient, shape-aware operations on implicit representations. We present three such applications, including a modification to sphere tracing that renders implicit representations with better convergence properties, a fast construction method for memory-efficient rigid-body collision proxies, and an efficient approximation of ambient occlusion that remains stable with respect to viewpoint variations.

110.Pointwise visual field estimation from optical coherence tomography in glaucoma: a structure-function analysis using deep learning ⬇️

Background/Aims: Standard Automated Perimetry (SAP) is the gold standard to monitor visual field (VF) loss in glaucoma management, but is prone to intra-subject variability. We developed and validated a deep learning (DL) regression model that estimates pointwise and overall VF loss from unsegmented optical coherence tomography (OCT) scans. Methods: Eight DL regression models were trained with various retinal imaging modalities: circumpapillary OCT at 3.5mm, 4.1mm, 4.7mm diameter, and scanning laser ophthalmoscopy (SLO) en face images to estimate mean deviation (MD) and 52 threshold values. This retrospective study used data from patients who underwent a complete glaucoma examination, including a reliable Humphrey Field Analyzer (HFA) 24-2 SITA Standard VF exam and a SPECTRALIS OCT scan using the Glaucoma Module Premium Edition. Results: A total of 1378 matched OCT-VF pairs of 496 patients (863 eyes) were included for training and evaluation of the DL models. Average sample MD was -7.53dB (from -33.8dB to +2.0dB). For 52 VF threshold values estimation, the circumpapillary OCT scan with the largest radius (4.7mm) achieved the best performance among all individual models (Pearson r=0.77, 95% CI=[0.72-0.82]). For MD, prediction averaging of OCT-trained models (3.5mm, 4.1mm, 4.7mm) resulted in a Pearson r of 0.78 [0.73-0.83] on the validation set and comparable performance on the test set (Pearson r=0.79 [0.75-0.82]). Conclusion: DL on unsegmented OCT scans accurately predicts pointwise and mean deviation of 24-2 VF in glaucoma patients. Automated VF from unsegmented OCT could be a solution for patients unable to produce reliable perimetry results.

111.Deep Neural Network-based Enhancement for Image and Video Streaming Systems: A Survey and Future Directions ⬇️

Internet-enabled smartphones and ultra-wide displays are transforming a variety of visual apps spanning from on-demand movies and 360° videos to video-conferencing and live streaming. However, robustly delivering visual content under fluctuating networking conditions on devices of diverse capabilities remains an open problem. In recent years, advances in the field of deep learning on tasks such as super-resolution and image enhancement have led to unprecedented performance in generating high-quality images from low-quality ones, a process we refer to as neural enhancement. In this paper, we survey state-of-the-art content delivery systems that employ neural enhancement as a key component in achieving both fast response time and high visual quality. We first present the components and architecture of existing content delivery systems, highlighting their challenges and motivating the use of neural enhancement models as a countermeasure. We then cover the deployment challenges of these models and analyze existing systems and their design decisions in efficiently overcoming these technical challenges. Additionally, we underline the key trends and common approaches across systems that target diverse use-cases. Finally, we present promising future directions based on the latest insights from deep learning research to further boost the quality of experience of content delivery systems.

112.Towards a Multi-purpose Robotic Nursing Assistant ⬇️

Robotic nursing aid is one of the heavily researched areas in robotics nowadays. Several robotic assistants exist that only focus on a specific function related to nurses assistance or functions related to patient aid. There is a need for a unified system that not only performs tasks that would assist nurses and reduce their burden but also perform tasks that help a patient. In recent times, due to the COVID-19 pandemic, there is also an increase in the need for robotic assistants that have teleoperation capabilities to provide better protection against the virus spread. To address these requirements, we propose a novel Multi-purpose Intelligent Nurse Aid (MINA) robotic system that is capable of providing walking assistance to the patients and perform teleoperation tasks with an easy-to-use and intuitive Graphical User Interface (GUI). This paper also presents preliminary results from the walking assistant task that improves upon the current state-of-the-art methods and shows the developed GUI for teleoperation.

113.Making EfficientNet More Efficient: Exploring Batch-Independent Normalization, Group Convolutions and Reduced Resolution Training ⬇️

Much recent research has been dedicated to improving the efficiency of training and inference for image classification. This effort has commonly focused on explicitly improving theoretical efficiency, often measured as ImageNet validation accuracy per FLOP. These theoretical savings have, however, proven challenging to achieve in practice, particularly on high-performance training accelerators.
In this work, we focus on improving the practical efficiency of the state-of-the-art EfficientNet models on a new class of accelerator, the Graphcore IPU. We do this by extending this family of models in the following ways: (i) generalising depthwise convolutions to group convolutions; (ii) adding proxy-normalized activations to match batch normalization performance with batch-independent statistics; (iii) reducing compute by lowering the training resolution and inexpensively fine-tuning at higher resolution. We find that these three methods improve the practical efficiency for both training and inference. Our code will be made available online.

114.Representation mitosis in wide neural networks ⬇️

Deep neural networks (DNNs) defy the classical bias-variance trade-off: adding parameters to a DNN that exactly interpolates its training data will typically improve its generalisation performance. Explaining the mechanism behind the benefit of such over-parameterisation is an outstanding challenge for deep learning theory. Here, we study the last layer representation of various deep architectures such as Wide-ResNets for image classification and find evidence for an underlying mechanism that we call representation mitosis: if the last hidden representation is wide enough, its neurons tend to split into groups which carry identical information, and differ from each other only by a statistically independent noise. Like in a mitosis process, the number of such groups, or ``clones'', increases linearly with the width of the layer, but only if the width is above a critical value. We show that a key ingredient to activate mitosis is continuing the training process until the training error is zero. Finally, we show that in one of the learning tasks we considered, a wide model with several automatically developed clones performs significantly better than a deep ensemble based on architectures in which the last layer has the same size as the clones.

115.Knowledge-aware Deep Framework for Collaborative Skin Lesion Segmentation and Melanoma Recognition ⬇️

Deep learning techniques have shown their superior performance in dermatologist clinical inspection. Nevertheless, melanoma diagnosis is still a challenging task due to the difficulty of incorporating the useful dermatologist clinical knowledge into the learning process. In this paper, we propose a novel knowledge-aware deep framework that incorporates some clinical knowledge into collaborative learning of two important melanoma diagnosis tasks, i.e., skin lesion segmentation and melanoma recognition. Specifically, to exploit the knowledge of morphological expressions of the lesion region and also the periphery region for melanoma identification, a lesion-based pooling and shape extraction (LPSE) scheme is designed, which transfers the structure information obtained from skin lesion segmentation into melanoma recognition. Meanwhile, to pass the skin lesion diagnosis knowledge from melanoma recognition to skin lesion segmentation, an effective diagnosis guided feature fusion (DGFF) strategy is designed. Moreover, we propose a recursive mutual learning mechanism that further promotes the inter-task cooperation, and thus iteratively improves the joint learning capability of the model for both skin lesion segmentation and melanoma recognition. Experimental results on two publicly available skin lesion datasets show the effectiveness of the proposed method for melanoma analysis.

116.Commutative Lie Group VAE for Disentanglement Learning ⬇️

We view disentanglement learning as discovering an underlying structure that equivariantly reflects the factorized variations shown in data. Traditionally, such a structure is fixed to be a vector space with data variations represented by translations along individual latent dimensions. We argue this simple structure is suboptimal since it requires the model to learn to discard the properties (e.g. different scales of changes, different levels of abstractness) of data variations, which is an extra work than equivariance learning. Instead, we propose to encode the data variations with groups, a structure not only can equivariantly represent variations, but can also be adaptively optimized to preserve the properties of data variations. Considering it is hard to conduct training on group structures, we focus on Lie groups and adopt a parameterization using Lie algebra. Based on the parameterization, some disentanglement learning constraints are naturally derived. A simple model named Commutative Lie Group VAE is introduced to realize the group-based disentanglement learning. Experiments show that our model can effectively learn disentangled representations without supervision, and can achieve state-of-the-art performance without extra constraints.

117.Continual Active Learning for Efficient Adaptation of Machine Learning Models to Changing Image Acquisition ⬇️

Imaging in clinical routine is subject to changing scanner protocols, hardware, or policies in a typically heterogeneous set of acquisition hardware. Accuracy and reliability of deep learning models suffer from those changes as data and targets become inconsistent with their initial static training set. Continual learning can adapt to a continuous data stream of a changing imaging environment. Here, we propose a method for continual active learning on a data stream of medical images. It recognizes shifts or additions of new imaging sources - domains -, adapts training accordingly, and selects optimal examples for labelling. Model training has to cope with a limited labelling budget, resembling typical real world scenarios. We demonstrate our method on T1-weighted magnetic resonance images from three different scanners with the task of brain age estimation. Results demonstrate that the proposed method outperforms naive active learning while requiring less manual labelling.

118.Zero-Shot Knowledge Distillation from a Decision-Based Black-Box Model ⬇️

Knowledge distillation (KD) is a successful approach for deep neural network acceleration, with which a compact network (student) is trained by mimicking the softmax output of a pre-trained high-capacity network (teacher). In tradition, KD usually relies on access to the training samples and the parameters of the white-box teacher to acquire the transferred knowledge. However, these prerequisites are not always realistic due to storage costs or privacy issues in real-world applications. Here we propose the concept of decision-based black-box (DB3) knowledge distillation, with which the student is trained by distilling the knowledge from a black-box teacher (parameters are not accessible) that only returns classes rather than softmax outputs. We start with the scenario when the training set is accessible. We represent a sample's robustness against other classes by computing its distances to the teacher's decision boundaries and use it to construct the soft label for each training sample. After that, the student can be trained via standard KD. We then extend this approach to a more challenging scenario in which even accessing the training data is not feasible. We propose to generate pseudo samples distinguished by the teacher's decision boundaries to the largest extent and construct soft labels for them, which are used as the transfer set. We evaluate our approaches on various benchmark networks and datasets and experiment results demonstrate their effectiveness. Codes are available at: this https URL.

119.Understand and Improve Contrastive Learning Methods for Visual Representation: A Review ⬇️

Traditional supervised learning methods are hitting a bottleneck because of their dependency on expensive manually labeled data and their weaknesses such as limited generalization ability and vulnerability to adversarial attacks. A promising alternative, self-supervised learning, as a type of unsupervised learning, has gained popularity because of its potential to learn effective data representations without manual labeling. Among self-supervised learning algorithms, contrastive learning has achieved state-of-the-art performance in several fields of research. This literature review aims to provide an up-to-date analysis of the efforts of researchers to understand the key components and the limitations of self-supervised learning.

120.Efficient Lottery Ticket Finding: Less Data is More ⬇️

The lottery ticket hypothesis (LTH) reveals the existence of winning tickets (sparse but critical subnetworks) for dense networks, that can be trained in isolation from random initialization to match the latter's accuracies. However, finding winning tickets requires burdensome computations in the train-prune-retrain process, especially on large-scale datasets (e.g., ImageNet), restricting their practical benefits. This paper explores a new perspective on finding lottery tickets more efficiently, by doing so only with a specially selected subset of data, called Pruning-Aware Critical set (PrAC set), rather than using the full training set. The concept of PrAC set was inspired by the recent observation, that deep networks have samples that are either hard to memorize during training, or easy to forget during pruning. A PrAC set is thus hypothesized to capture those most challenging and informative examples for the dense model. We observe that a high-quality winning ticket can be found with training and pruning the dense network on the very compact PrAC set, which can substantially save training iterations for the ticket finding process. Extensive experiments validate our proposal across diverse datasets and network architectures. Specifically, on CIFAR-10, CIFAR-100, and Tiny ImageNet, we locate effective PrAC sets at 35.32%~78.19% of their training set sizes. On top of them, we can obtain the same competitive winning tickets for the corresponding dense networks, yet saving up to 82.85%~92.77%, 63.54%~74.92%, and 76.14%~86.56% training iterations, respectively. Crucially, we show that a PrAC set found is reusable across different network architectures, which can amortize the extra cost of finding PrAC sets, yielding a practical regime for efficient lottery ticket finding.

121.CAPE: Encoding Relative Positions with Continuous Augmented Positional Embeddings ⬇️

Without positional information, attention-based transformer neural networks are permutation-invariant. Absolute or relative positional embeddings are the most popular ways to feed transformer models positional information. Absolute positional embeddings are simple to implement, but suffer from generalization issues when evaluating on sequences of different length than those seen at training time. Relative positions are more robust to length change, but are more complex to implement and yield inferior model throughput. In this paper, we propose an augmentation-based approach (CAPE) for absolute positional embeddings, which keeps the advantages of both absolute (simplicity and speed) and relative position embeddings (better generalization). In addition, our empirical evaluation on state-of-the-art models in machine translation, image and speech recognition demonstrates that CAPE leads to better generalization performance as well as increased stability with respect to training hyper-parameters.

122.End-to-End Neuro-Symbolic Architecture for Image-to-Image Reasoning Tasks ⬇️

Neural models and symbolic algorithms have recently been combined for tasks requiring both perception and reasoning. Neural models ground perceptual input into a conceptual vocabulary, on which a classical reasoning algorithm is applied to generate output. A key limitation is that such neural-to-symbolic models can only be trained end-to-end for tasks where the output space is symbolic. In this paper, we study neural-symbolic-neural models for reasoning tasks that require a conversion from an image input (e.g., a partially filled sudoku) to an image output (e.g., the image of the completed sudoku). While designing such a three-step hybrid architecture may be straightforward, the key technical challenge is end-to-end training -- how to backpropagate without intermediate supervision through the symbolic component. We propose NSNnet, an architecture that combines an image reconstruction loss with a novel output encoder to generate a supervisory signal, develops update algorithms that leverage policy gradient methods for supervision, and optimizes loss using a novel subsampling heuristic. We experiment on problem settings where symbolic algorithms are easily specified: a visual maze solving task and a visual Sudoku solver where the supervision is in image form. Experiments show high accuracy with significantly less data compared to purely neural approaches.

123.Asymmetric Loss Functions for Learning with Noisy Labels ⬇️

Robust loss functions are essential for training deep neural networks with better generalization power in the presence of noisy labels. Symmetric loss functions are confirmed to be robust to label noise. However, the symmetric condition is overly restrictive. In this work, we propose a new class of loss functions, namely \textit{asymmetric loss functions}, which are robust to learning with noisy labels for various types of noise. We investigate general theoretical properties of asymmetric loss functions, including classification calibration, excess risk bound, and noise tolerance. Meanwhile, we introduce the asymmetry ratio to measure the asymmetry of a loss function. The empirical results show that a higher ratio would provide better noise tolerance. Moreover, we modify several commonly-used loss functions and establish the necessary and sufficient conditions for them to be asymmetric. Experimental results on benchmark datasets demonstrate that asymmetric loss functions can outperform state-of-the-art methods. The code is available at \href{this https URL}{this https URL}

124.Preservation of the Global Knowledge by Not-True Self Knowledge Distillation in Federated Learning ⬇️

In Federated Learning (FL), a strong global model is collaboratively learned by aggregating the clients' locally trained models. Although this allows no need to access clients' data directly, the global model's convergence often suffers from data heterogeneity. This paper suggests that forgetting could be the bottleneck of global convergence. We observe that fitting on biased local distribution shifts the feature on global distribution and results in forgetting of global knowledge. We consider this phenomenon as an analogy to Continual Learning, which also faces catastrophic forgetting when fitted on the new task distribution. Based on our findings, we hypothesize that tackling down the forgetting in local training relives the data heterogeneity problem. To this end, we propose a simple yet effective framework Federated Local Self-Distillation (FedLSD), which utilizes the global knowledge on locally available data. By following the global perspective on local data, FedLSD encourages the learned features to preserve global knowledge and have consistent views across local models, thus improving convergence without compromising data privacy. Under our framework, we further extend FedLSD to FedLS-NTD, which only considers the not-true class signals to compensate noisy prediction of the global model. We validate that both FedLSD and FedLS-NTD significantly improve the performance in standard FL benchmarks in various setups, especially in the extreme data heterogeneity cases.

125.Brain Age Estimation From MRI Using Cascade Networks with Ranking Loss ⬇️

Chronological age of healthy people is able to be predicted accurately using deep neural networks from neuroimaging data, and the predicted brain age could serve as a biomarker for detecting aging-related diseases. In this paper, a novel 3D convolutional network, called two-stage-age-network (TSAN), is proposed to estimate brain age from T1-weighted MRI data. Compared with existing methods, TSAN has the following improvements. First, TSAN uses a two-stage cascade network architecture, where the first-stage network estimates a rough brain age, then the second-stage network estimates the brain age more accurately from the discretized brain age by the first-stage network. Second, to our knowledge, TSAN is the first work to apply novel ranking losses in brain age estimation, together with the traditional mean square error (MSE) loss. Third, densely connected paths are used to combine feature maps with different scales. The experiments with $6586$ MRIs showed that TSAN could provide accurate brain age estimation, yielding mean absolute error (MAE) of $2.428$ and Pearson's correlation coefficient (PCC) of $0.985$, between the estimated and chronological ages. Furthermore, using the brain age gap between brain age and chronological age as a biomarker, Alzheimer's disease (AD) and Mild Cognitive Impairment (MCI) can be distinguished from healthy control (HC) subjects by support vector machine (SVM). Classification AUC in AD/HC and MCI/HC was $0.904$ and $0.823$, respectively. It showed that brain age gap is an effective biomarker associated with risk of dementia, and has potential for early-stage dementia risk screening. The codes and trained models have been released on GitHub: this https URL.

126.DAMSL: Domain Agnostic Meta Score-based Learning ⬇️

In this paper, we propose Domain Agnostic Meta Score-based Learning (DAMSL), a novel, versatile and highly effective solution that delivers significant out-performance over state-of-the-art methods for cross-domain few-shot learning. We identify key problems in previous meta-learning methods over-fitting to the source domain, and previous transfer-learning methods under-utilizing the structure of the support set. The core idea behind our method is that instead of directly using the scores from a fine-tuned feature encoder, we use these scores to create input coordinates for a domain agnostic metric space. A graph neural network is applied to learn an embedding and relation function over these coordinates to process all information contained in the score distribution of the support set. We test our model on both established CD-FSL benchmarks and new domains and show that our method overcomes the limitations of previous meta-learning and transfer-learning methods to deliver substantial improvements in accuracy across both smaller and larger domain shifts.

127.Local Disentanglement in Variational Auto-Encoders Using Jacobian $L_1$ Regularization ⬇️

There have been many recent advances in representation learning; however, unsupervised representation learning can still struggle with model identification issues. Variational Auto-Encoders (VAEs) and their extensions such as $\beta$-VAEs have been shown to locally align latent variables with PCA directions, which can help to improve model disentanglement under some conditions. Borrowing inspiration from Independent Component Analysis (ICA) and sparse coding, we propose applying an $L_1$ loss to the VAE's generative Jacobian during training to encourage local latent variable alignment with independent factors of variation in the data. We demonstrate our results on a variety of datasets, giving qualitative and quantitative results using information theoretic and modularity measures that show our added $L_1$ cost encourages local axis alignment of the latent representation with individual factors of variation.

128.A Deep Variational Bayesian Framework for Blind Image Deblurring ⬇️

Blind image deblurring is an important yet very challenging problem in low-level vision. Traditional optimization based methods generally formulate this task as a maximum-a-posteriori estimation or variational inference problem, whose performance highly relies on the handcraft priors for both the latent image and the blur kernel. In contrast, recent deep learning methods generally learn, from a large collection of training images, deep neural networks (DNNs) directly mapping the blurry image to the clean one or to the blur kernel, paying less attention to the physical degradation process of the blurry image. In this paper, we present a deep variational Bayesian framework for blind image deblurring. Under this framework, the posterior of the latent clean image and blur kernel can be jointly estimated in an amortized inference fashion with DNNs, and the involved inference DNNs can be trained by fully considering the physical blur model, together with the supervision of data driven priors for the clean image and blur kernel, which is naturally led to by the evidence lower bound objective. Comprehensive experiments are conducted to substantiate the effectiveness of the proposed framework. The results show that it can not only achieve a promising performance with relatively simple networks, but also enhance the performance of existing DNNs for deblurring.

129.AOSLO-net: A deep learning-based method for automatic segmentation of retinal microaneurysms from adaptive optics scanning laser ophthalmoscope images ⬇️

Adaptive optics scanning laser ophthalmoscopy (AOSLO) provides real-time retinal images with high resolution down to 2 $\mu m$. This technique enables detection of the morphologies of individual microaneurysms (MAs), which are one of the earliest signs of diabetic retinopathy (DR), a frequent complication of diabetes that can lead to visual impairment and blindness. In contrast to previous automatic models developed for MA detection on standard fundus photographs, currently there is no high throughput image protocol available for automatic analysis of AOSLO photographs. To address this urgency, we introduce AOSLO-net, a deep neural network framework with customized training policy, including preprocessing, data augmentation and transfer learning, to automatically segment MAs from AOSLO images. We evaluate the performance of AOSLO-net using 87 DR AOSLO images demonstrating very accurate MA detection and segmentation, leading to correct MA morphological classification, while outperforming the state-of-the-art both in accuracy and cost.

130.Principle Bit Analysis: Autoencoding with Schur-Concave Loss ⬇️

We consider a linear autoencoder in which the latent variables are quantized, or corrupted by noise, and the constraint is Schur-concave in the set of latent variances. Although finding the optimal encoder/decoder pair for this setup is a nonconvex optimization problem, we show that decomposing the source into its principal components is optimal. If the constraint is strictly Schur-concave and the empirical covariance matrix has only simple eigenvalues, then any optimal encoder/decoder must decompose the source in this way. As one application, we consider a strictly Schur-concave constraint that estimates the number of bits needed to represent the latent variables under fixed-rate encoding, a setup that we call \emph{Principal Bit Analysis (PBA)}. This yields a practical, general-purpose, fixed-rate compressor that outperforms existing algorithms. As a second application, we show that a prototypical autoencoder-based variable-rate compressor is guaranteed to decompose the source into its principal components.

131.Learnable Fourier Features for Multi-DimensionalSpatial Positional Encoding ⬇️

Attentional mechanisms are order-invariant. Positional encoding is a crucial component to allow attention-based deep model architectures such as Transformer to address sequences or images where the position of information matters. In this paper, we propose a novel positional encoding method based on learnable Fourier features. Instead of hard-coding each position as a token or a vector, we represent each position, which can be multi-dimensional, as a trainable encoding based on learnable Fourier feature mapping, modulated with a multi-layer perceptron. The representation is particularly advantageous for a spatial multi-dimensional position, e.g., pixel positions on an image, where $L_2$ distances or more complex positional relationships need to be captured. Our experiments based on several public benchmark tasks show that our learnable Fourier feature representation for multi-dimensional positional encoding outperforms existing methods by both improving the accuracy and allowing faster convergence.

132.SketchGen: Generating Constrained CAD Sketches ⬇️

Computer-aided design (CAD) is the most widely used modeling approach for technical design. The typical starting point in these designs is 2D sketches which can later be extruded and combined to obtain complex three-dimensional assemblies. Such sketches are typically composed of parametric primitives, such as points, lines, and circular arcs, augmented with geometric constraints linking the primitives, such as coincidence, parallelism, or orthogonality. Sketches can be represented as graphs, with the primitives as nodes and the constraints as edges. Training a model to automatically generate CAD sketches can enable several novel workflows, but is challenging due to the complexity of the graphs and the heterogeneity of the primitives and constraints. In particular, each type of primitive and constraint may require a record of different size and parameter types. We propose SketchGen as a generative model based on a transformer architecture to address the heterogeneity problem by carefully designing a sequential language for the primitives and constraints that allows distinguishing between different primitive or constraint types and their parameters, while encouraging our model to re-use information across related parameters, encoding shared structure. A particular highlight of our work is the ability to produce primitives linked via constraints that enables the final output to be further regularized via a constraint solver. We evaluate our model by demonstrating constraint prediction for given sets of primitives and full sketch generation from scratch, showing that our approach significantly out performs the state-of-the-art in CAD sketch generation.

133.Real Time Video based Heart and Respiration Rate Monitoring ⬇️

In recent years, research about monitoring vital signs by smartphones grows significantly. There are some special sensors like Electrocardiogram (ECG) and Photoplethysmographic (PPG) to detect heart rate (HR) and respiration rate (RR). Smartphone cameras also can measure HR by detecting and processing imaging Photoplethysmographic (iPPG) signals from the video of a user's face. Indeed, the variation in the intensity of the green channel can be measured by the iPPG signals of the video. This study aimed to provide a method to extract heart rate and respiration rate using the video of individuals' faces. The proposed method is based on measuring fluctuations in the Hue, and can therefore extract both HR and RR from the video of a user's face. The proposed method is evaluated by performing on 25 healthy individuals. For each subject, 20 seconds video of his/her face is recorded. Results show that the proposed approach of measuring iPPG using Hue gives more accurate rates than the Green channel.