20220324.md

ArXiv cs.CV --Thu, 24 Mar 2022

1.Unsupervised Salient Object Detection with Spectral Cluster Voting ⬇️

In this paper, we tackle the challenging task of unsupervised salient object detection (SOD) by leveraging spectral clustering on self-supervised features. We make the following contributions: (i) We revisit spectral clustering and demonstrate its potential to group the pixels of salient objects; (ii) Given mask proposals from multiple applications of spectral clustering on image features computed from various self-supervised models, e.g., MoCov2, SwAV, DINO, we propose a simple but effective winner-takes-all voting mechanism for selecting the salient masks, leveraging object priors based on framing and distinctiveness; (iii) Using the selected object segmentation as pseudo groundtruth masks, we train a salient object detector, dubbed SelfMask, which outperforms prior approaches on three unsupervised SOD benchmarks. Code is publicly available at this https URL.

2.A Hybrid Mesh-neural Representation for 3D Transparent Object Reconstruction ⬇️

We propose a novel method to reconstruct the 3D shapes of transparent objects using hand-held captured images under natural light conditions. It combines the advantage of explicit mesh and multi-layer perceptron (MLP) network, a hybrid representation, to simplify the capture setting used in recent contributions. After obtaining an initial shape through the multi-view silhouettes, we introduce surface-based local MLPs to encode the vertex displacement field (VDF) for the reconstruction of surface details. The design of local MLPs allows to represent the VDF in a piece-wise manner using two layer MLP networks, which is beneficial to the optimization algorithm. Defining local MLPs on the surface instead of the volume also reduces the searching space. Such a hybrid representation enables us to relax the ray-pixel correspondences that represent the light path constraint to our designed ray-cell correspondences, which significantly simplifies the implementation of single-image based environment matting algorithm. We evaluate our representation and reconstruction algorithm on several transparent objects with ground truth models. Our experiments show that our method can produce high-quality reconstruction results superior to state-of-the-art methods using a simplified data acquisition setup.

3.StructToken : Rethinking Semantic Segmentation with Structural Prior ⬇️

In this paper, we present structure token (StructToken), a new paradigm for semantic segmentation. From a perspective on semantic segmentation as per-pixel classification, the previous deep learning-based methods learn the per-pixel representation first through an encoder and a decoder head and then classify each pixel representation to a specific category to obtain the semantic masks. Differently, we propose a structure-aware algorithm that takes structural information as prior to predict semantic masks directly without per-pixel classification. Specifically, given an input image, the learnable structure token interacts with the image representations to reason the final semantic masks. Three interaction approaches are explored and the results not only outperform the state-of-the-art methods but also contain more structural information. Experiments are conducted on three widely used datasets including ADE20k, Cityscapes, and COCO-Stuff 10K. We hope that structure token could serve as an alternative for semantic segmentation and inspire future research.

4.VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training ⬇️

Pre-training video transformers on extra large-scale datasets is generally required to achieve premier performance on relatively small datasets. In this paper, we show that video masked autoencoders (VideoMAE) are data-efficient learners for self-supervised video pre-training (SSVP). We are inspired by the recent ImageMAE and propose customized video tube masking and reconstruction. These simple designs turn out to be effective for overcoming information leakage caused by the temporal correlation during video reconstruction. We obtain three important findings on SSVP: (1) An extremely high proportion of masking ratio (i.e., 90% to 95%) still yields favorable performance of VideoMAE. The temporally redundant video content enables higher masking ratio than that of images. (2) VideoMAE achieves impressive results on very small datasets (i.e., around 3k-4k videos) without using any extra data. This is partially ascribed to the challenging task of video reconstruction to enforce high-level structure learning. (3) VideoMAE shows that data quality is more important than data quantity for SSVP. Domain shift between pre-training and target datasets are important issues in SSVP. Notably, our VideoMAE with the vanilla ViT backbone can achieve 83.9% on Kinects-400, 75.3% on Something-Something V2, 90.8% on UCF101, and 61.1% on HMDB51 without using any extra data. Code will be released at this https URL.

5.NeuMan: Neural Human Radiance Field from a Single Video ⬇️

Photorealistic rendering and reposing of humans is important for enabling augmented reality experiences. We propose a novel framework to reconstruct the human and the scene that can be rendered with novel human poses and views from just a single in-the-wild video. Given a video captured by a moving camera, we train two NeRF models: a human NeRF model and a scene NeRF model. To train these models, we rely on existing methods to estimate the rough geometry of the human and the scene. Those rough geometry estimates allow us to create a warping field from the observation space to the canonical pose-independent space, where we train the human model in. Our method is able to learn subject specific details, including cloth wrinkles and accessories, from just a 10 seconds video clip, and to provide high quality renderings of the human under novel poses, from novel views, together with the background.

6.Your "Attention" Deserves Attention: A Self-Diversified Multi-Channel Attention for Facial Action Analysis ⬇️

Visual attention has been extensively studied for learning fine-grained features in both facial expression recognition (FER) and Action Unit (AU) detection. A broad range of previous research has explored how to use attention modules to localize detailed facial parts (e,g. facial action units), learn discriminative features, and learn inter-class correlation. However, few related works pay attention to the robustness of the attention module itself. Through experiments, we found neural attention maps initialized with different feature maps yield diverse representations when learning to attend the identical Region of Interest (ROI). In other words, similar to general feature learning, the representational quality of attention maps also greatly affects the performance of a model, which means unconstrained attention learning has lots of randomnesses. This uncertainty lets conventional attention learning fall into sub-optimal. In this paper, we propose a compact model to enhance the representational and focusing power of neural attention maps and learn the "inter-attention" correlation for refined attention maps, which we term the "Self-Diversified Multi-Channel Attention Network (SMA-Net)". The proposed method is evaluated on two benchmark databases (BP4D and DISFA) for AU detection and four databases (CK+, MMI, BU-3DFE, and BP4D+) for facial expression recognition. It achieves superior performance compared to the state-of-the-art methods.

7.DynamicEarthNet: Daily Multi-Spectral Satellite Dataset for Semantic Change Segmentation ⬇️

Earth observation is a fundamental tool for monitoring the evolution of land use in specific areas of interest. Observing and precisely defining change, in this context, requires both time-series data and pixel-wise segmentations. To that end, we propose the DynamicEarthNet dataset that consists of daily, multi-spectral satellite observations of 75 selected areas of interest distributed over the globe with imagery from Planet Labs. These observations are paired with pixel-wise monthly semantic segmentation labels of 7 land use and land cover (LULC) classes. DynamicEarthNet is the first dataset that provides this unique combination of daily measurements and high-quality labels. In our experiments, we compare several established baselines that either utilize the daily observations as additional training data (semi-supervised learning) or multiple observations at once (spatio-temporal learning) as a point of reference for future research. Finally, we propose a new evaluation metric SCS that addresses the specific challenges associated with time-series semantic change segmentation. The data is available at: this https URL.

8.Multi-label Transformer for Action Unit Detection ⬇️

Action Unit (AU) Detection is the branch of affective computing that aims at recognizing unitary facial muscular movements. It is key to unlock unbiaised computational face representations and has therefore aroused great interest in the past few years. One of main obstacles toward building efficient deep learning based AU detection system facial images database annotated by AU experts. In that extent the ABAW challenge paves the way toward better AU detection as it involves a ~2M frames AU annotated dataset. In this paper, we present our submission to the ABAW3 challenge. In a nutshell, we applied a multi-label detection transformer that leverage multi-head attention to learn which part of the face image is the most relevant to predict each AU.

9.Refine-Net: Normal Refinement Neural Network for Noisy Point Clouds ⬇️

Point normal, as an intrinsic geometric property of 3D objects, not only serves conventional geometric tasks such as surface consolidation and reconstruction, but also facilitates cutting-edge learning-based techniques for shape analysis and generation. In this paper, we propose a normal refinement network, called Refine-Net, to predict accurate normals for noisy point clouds. Traditional normal estimation wisdom heavily depends on priors such as surface shapes or noise distributions, while learning-based solutions settle for single types of hand-crafted features. Differently, our network is designed to refine the initial normal of each point by extracting additional information from multiple feature representations. To this end, several feature modules are developed and incorporated into Refine-Net by a novel connection module. Besides the overall network architecture of Refine-Net, we propose a new multi-scale fitting patch selection scheme for the initial normal estimation, by absorbing geometry domain knowledge. Also, Refine-Net is a generic normal estimation framework: 1) point normals obtained from other methods can be further refined, and 2) any feature module related to the surface geometric structures can be potentially integrated into the framework. Qualitative and quantitative evaluations demonstrate the clear superiority of Refine-Net over the state-of-the-arts on both synthetic and real-scanned datasets. Our code is available at this https URL.

10.CroMo: Cross-Modal Learning for Monocular Depth Estimation ⬇️

Learning-based depth estimation has witnessed recent progress in multiple directions; from self-supervision using monocular video to supervised methods offering highest accuracy. Complementary to supervision, further boosts to performance and robustness are gained by combining information from multiple signals. In this paper we systematically investigate key trade-offs associated with sensor and modality design choices as well as related model training strategies. Our study leads us to a new method, capable of connecting modality-specific advantages from polarisation, Time-of-Flight and structured-light inputs. We propose a novel pipeline capable of estimating depth from monocular polarisation for which we evaluate various training signals. The inversion of differentiable analytic models thereby connects scene geometry with polarisation and ToF signals and enables self-supervised and cross-modal learning. In the absence of existing multimodal datasets, we examine our approach with a custom-made multi-modal camera rig and collect CroMo; the first dataset to consist of synchronized stereo polarisation, indirect ToF and structured-light depth, captured at video rates. Extensive experiments on challenging video scenes confirm both qualitative and quantitative pipeline advantages where we are able to outperform competitive monocular depth estimation method.

11.A Deep Learning Framework to Reconstruct Face under Mask ⬇️

While deep learning-based image reconstruction methods have shown significant success in removing objects from pictures, they have yet to achieve acceptable results for attributing consistency to gender, ethnicity, expression, and other characteristics like the topological structure of the face. The purpose of this work is to extract the mask region from a masked image and rebuild the area that has been detected. This problem is complex because (i) it is difficult to determine the gender of an image hidden behind a mask, which causes the network to become confused and reconstruct the male face as a female or vice versa; (ii) we may receive images from multiple angles, making it extremely difficult to maintain the actual shape, topological structure of the face and a natural image; and (iii) there are problems with various mask forms because, in some cases, the area of the mask cannot be anticipated precisely; certain parts of the mask remain on the face after completion. To solve this complex task, we split the problem into three phases: landmark detection, object detection for the targeted mask area, and inpainting the addressed mask region. To begin, to solve the first problem, we have used gender classification, which detects the actual gender behind a mask, then we detect the landmark of the masked facial image. Second, we identified the non-face item, i.e., the mask, and used the Mask R-CNN network to create the binary mask of the observed mask area. Thirdly, we developed an inpainting network that uses anticipated landmarks to create realistic images. To segment the mask, this article uses a mask R-CNN and offers a binary segmentation map for identifying the mask area. Additionally, we generated the image utilizing landmarks as structural guidance through a GAN-based network. The studies presented in this paper use the FFHQ and CelebA datasets.

12.Adaptively Re-weighting Multi-Loss Untrained Transformer for Sparse-View Cone-Beam CT Reconstruction ⬇️

Cone-Beam Computed Tomography (CBCT) has been proven useful in diagnosis, but how to shorten scanning time with lower radiation dosage and how to efficiently reconstruct 3D image remain as the main issues for clinical practice. The recent development of tomographic image reconstruction on sparse-view measurements employs deep neural networks in a supervised way to tackle such issues, whereas the success of model training requires quantity and quality of the given paired measurements/images. We propose a novel untrained Transformer to fit the CBCT inverse solver without training data. It is mainly comprised of an untrained 3D Transformer of billions of network weights and a multi-level loss function with variable weights. Unlike conventional deep neural networks (DNNs), there is no requirement of training steps in our approach. Upon observing the hardship of optimising Transformer, the variable weights within the loss function are designed to automatically update together with the iteration process, ultimately stabilising its optimisation. We evaluate the proposed approach on two publicly available datasets: SPARE and Walnut. The results show a significant performance improvement on image quality metrics with streak artefact reduction in the visualisation. We also provide a clinical report by an experienced radiologist to assess our reconstructed images in a diagnosis point of view. The source code and the optimised models are available from the corresponding author on request at the moment.

13.3D Adapted Random Forest Vision (3DARFV) for Untangling Heterogeneous-Fabric Exceeding Deep Learning Semantic Segmentation Efficiency at the Utmost Accuracy ⬇️

Planetary exploration depends heavily on 3D image data to characterize the static and dynamic properties of the rock and environment. Analyzing 3D images requires many computations, causing efficiency to suffer lengthy processing time alongside large energy consumption. High-Performance Computing (HPC) provides apparent efficiency at the expense of energy consumption. However, for remote explorations, the conveyed surveillance and the robotized sensing need faster data analysis with ultimate accuracy to make real-time decisions. In such environments, access to HPC and energy is limited. Therefore, we realize that reducing the number of computations to optimal and maintaining the desired accuracy leads to higher efficiency. This paper demonstrates the semantic segmentation capability of a probabilistic decision tree algorithm, 3D Adapted Random Forest Vision (3DARFV), exceeding deep learning algorithm efficiency at the utmost accuracy.

14.Activation-Based Sampling for Pixel- to Image-Level Aggregation in Weakly-Supervised Segmentation ⬇️

Classification networks can be used to localize and segment objects in images by means of class activation maps (CAMs). However, without pixel-level annotations, they are known to (1) mainly focus on discriminative regions, and (2) to produce diffuse CAMs without well-defined prediction contours. In this work, we approach both problems with two contributions for improving CAM learning. First, we incorporate importance sampling based on the class-wise probability mass function induced by the CAMs to produce stochastic image-level class predictions. This results in CAMs which activate over a larger extent of the objects. Second, we formulate a feature similarity loss term which aims to match the prediction contours with edges in the image. As a third contribution, we conduct experiments on the PASCAL VOC and MS-COCO benchmark datasets to demonstrate that these modifications significantly increase the performance in terms of contour accuracy, while being comparable to current state-of-the-art methods in terms of region similarity.

15.MT-UDA: Towards Unsupervised Cross-modality Medical Image Segmentation with Limited Source Labels ⬇️

The success of deep convolutional neural networks (DCNNs) benefits from high volumes of annotated data. However, annotating medical images is laborious, expensive, and requires human expertise, which induces the label scarcity problem. Especially when encountering the domain shift, the problem becomes more serious. Although deep unsupervised domain adaptation (UDA) can leverage well-established source domain annotations and abundant target domain data to facilitate cross-modality image segmentation and also mitigate the label paucity problem on the target domain, the conventional UDA methods suffer from severe performance degradation when source domain annotations are scarce. In this paper, we explore a challenging UDA setting - limited source domain annotations. We aim to investigate how to efficiently leverage unlabeled data from the source and target domains with limited source annotations for cross-modality image segmentation. To achieve this, we propose a new label-efficient UDA framework, termed MT-UDA, in which the student model trained with limited source labels learns from unlabeled data of both domains by two teacher models respectively in a semi-supervised manner. More specifically, the student model not only distills the intra-domain semantic knowledge by encouraging prediction consistency but also exploits the inter-domain anatomical information by enforcing structural consistency. Consequently, the student model can effectively integrate the underlying knowledge beneath available data resources to mitigate the impact of source label scarcity and yield improved cross-modality segmentation performance. We evaluate our method on MM-WHS 2017 dataset and demonstrate that our approach outperforms the state-of-the-art methods by a large margin under the source-label scarcity scenario.

16.SMEMO: Social Memory for Trajectory Forecasting ⬇️

Effective modeling of human interactions is of utmost importance when forecasting behaviors such as future trajectories. Each individual, with its motion, influences surrounding agents since everyone obeys to social non-written rules such as collision avoidance or group following. In this paper we model such interactions, which constantly evolve through time, by looking at the problem from an algorithmic point of view, i.e. as a data manipulation task. We present a neural network based on an end-to-end trainable working memory, which acts as an external storage where information about each agent can be continuously written, updated and recalled. We show that our method is capable of learning explainable cause-effect relationships between motions of different agents, obtaining state-of-the-art results on multiple trajectory forecasting datasets.

17.An Attention-based Method for Action Unit Detection at the 3rd ABAW Competition ⬇️

Facial Action Coding System is an approach for modeling the complexity of human emotional expression. Automatic action unit (AU) detection is a crucial research area in human-computer interaction. This paper describes our submission to the third Affective Behavior Analysis in-the-wild (ABAW) competition 2022. We proposed a method for detecting facial action units in the video. At the first stage, a lightweight CNN-based feature extractor is employed to extract the feature map from each video frame. Then, an attention module is applied to refine the attention map. The attention encoded vector is derived using a weighted sum of the feature map and the attention scores later. Finally, the sigmoid function is used at the output layer to make the prediction suitable for multi-label AUs detection. We achieved a macro F1 score of 0.48 on the ABAW challenge validation set compared to 0.39 from the baseline model.

18.On the (Limited) Generalization of MasterFace Attacks and Its Relation to the Capacity of Face Representations ⬇️

A MasterFace is a face image that can successfully match against a large portion of the population. Since their generation does not require access to the information of the enrolled subjects, MasterFace attacks represent a potential security risk for widely-used face recognition systems. Previous works proposed methods for generating such images and demonstrated that these attacks can strongly compromise face recognition. However, previous works followed evaluation settings consisting of older recognition models, limited cross-dataset and cross-model evaluations, and the use of low-scale testing data. This makes it hard to state the generalizability of these attacks. In this work, we comprehensively analyse the generalizability of MasterFace attacks in empirical and theoretical investigations. The empirical investigations include the use of six state-of-the-art FR models, cross-dataset and cross-model evaluation protocols, and utilizing testing datasets of significantly higher size and variance. The results indicate a low generalizability when MasterFaces are training on a different face recognition model than the one used for testing. In these cases, the attack performance is similar to zero-effort imposter attacks. In the theoretical investigations, we define and estimate the face capacity and the maximum MasterFace coverage under the assumption that identities in the face space are well separated. The current trend of increasing the fairness and generalizability in face recognition indicates that the vulnerability of future systems might further decrease. We conclude that MasterFaces should not be seen as a threat to face recognition systems but, on the contrary, seen as a tool to understand and enhance the robustness of face recognition models.

19.Transformer-based Multimodal Information Fusion for Facial Expression Analysis ⬇️

Facial expression analysis has been a crucial research problem in the computer vision area. With the recent development of deep learning techniques and large-scale in-the-wild annotated datasets, facial expression analysis is now aimed at challenges in real world settings. In this paper, we introduce our submission to CVPR2022 Competition on Affective Behavior Analysis in-the-wild (ABAW) that defines four competition tasks, including expression classification, action unit detection, valence-arousal estimation, and a multi-task-learning. The available multimodal information consist of spoken words, speech prosody, and visual expression in videos. Our work proposes four unified transformer-based network frameworks to create the fusion of the above multimodal information. The preliminary results on the official Aff-Wild2 dataset are reported and demonstrate the effectiveness of our proposed method.

20.Hyper-Spectral Imaging for Overlapping Plastic Flakes Segmentation ⬇️

Given the hyper-spectral imaging unique potentials in grasping the polymer characteristics of different materials, it is commonly used in sorting procedures. In a practical plastic sorting scenario, multiple plastic flakes may overlap which depending on their characteristics, the overlap can be reflected in their spectral signature. In this work, we use hyper-spectral imaging for the segmentation of three types of plastic flakes and their possible overlapping combinations. We propose an intuitive and simple multi-label encoding approach, bitfield encoding, to account for the overlapping regions. With our experiments, we show that the bitfield encoding improves over the baseline single-label approach and we further demonstrate its potential in predicting multiple labels for overlapping classes even when the model is only trained with non-overlapping classes.

21.Robust Text Line Detection in Historical Documents: Learning and Evaluation Methods ⬇️

Text line segmentation is one of the key steps in historical document understanding. It is challenging due to the variety of fonts, contents, writing styles and the quality of documents that have degraded through the years.
In this paper, we address the limitations that currently prevent people from building line segmentation models with a high generalization capacity. We present a study conducted using three state-of-the-art systems Doc-UFCN, dhSegment and ARU-Net and show that it is possible to build generic models trained on a wide variety of historical document datasets that can correctly segment diverse unseen pages. This paper also highlights the importance of the annotations used during training: each existing dataset is annotated differently. We present a unification of the annotations and show its positive impact on the final text recognition results. In this end, we present a complete evaluation strategy using standard pixel-level metrics, object-level ones and introducing goal-oriented metrics.

22.How Do You Do It? Fine-Grained Action Understanding with Pseudo-Adverbs ⬇️

We aim to understand how actions are performed and identify subtle differences, such as 'fold firmly' vs. 'fold gently'. To this end, we propose a method which recognizes adverbs across different actions. However, such fine-grained annotations are difficult to obtain and their long-tailed nature makes it challenging to recognize adverbs in rare action-adverb compositions. Our approach therefore uses semi-supervised learning with multiple adverb pseudo-labels to leverage videos with only action labels. Combined with adaptive thresholding of these pseudo-adverbs we are able to make efficient use of the available data while tackling the long-tailed distribution. Additionally, we gather adverb annotations for three existing video retrieval datasets, which allows us to introduce the new tasks of recognizing adverbs in unseen action-adverb compositions and unseen domains. Experiments demonstrate the effectiveness of our method, which outperforms prior work in recognizing adverbs and semi-supervised works adapted for adverb recognition. We also show how adverbs can relate fine-grained actions.

23.Towards Semi-Supervised Deep Facial Expression Recognition with An Adaptive Confidence Margin ⬇️

Only parts of unlabeled data are selected to train models for most semi-supervised learning methods, whose confidence scores are usually higher than the pre-defined threshold (i.e., the confidence margin). We argue that the recognition performance should be further improved by making full use of all unlabeled data. In this paper, we learn an Adaptive Confidence Margin (Ada-CM) to fully leverage all unlabeled data for semi-supervised deep facial expression recognition. All unlabeled samples are partitioned into two subsets by comparing their confidence scores with the adaptively learned confidence margin at each training epoch: (1) subset I including samples whose confidence scores are no lower than the margin; (2) subset II including samples whose confidence scores are lower than the margin. For samples in subset I, we constrain their predictions to match pseudo labels. Meanwhile, samples in subset II participate in the feature-level contrastive objective to learn effective facial expression features. We extensively evaluate Ada-CM on four challenging datasets, showing that our method achieves state-of-the-art performance, especially surpassing fully-supervised baselines in a semi-supervised manner. Ablation study further proves the effectiveness of our method. The source code is available at this https URL.

24.Real-time Object Detection for Streaming Perception ⬇️

Autonomous driving requires the model to perceive the environment and (re)act within a low latency for safety. While past works ignore the inevitable changes in the environment after processing, streaming perception is proposed to jointly evaluate the latency and accuracy into a single metric for video online perception. In this paper, instead of searching trade-offs between accuracy and speed like previous works, we point out that endowing real-time models with the ability to predict the future is the key to dealing with this problem. We build a simple and effective framework for streaming perception. It equips a novel DualFlow Perception module (DFP), which includes dynamic and static flows to capture the moving trend and basic detection feature for streaming prediction. Further, we introduce a Trend-Aware Loss (TAL) combined with a trend factor to generate adaptive weights for objects with different moving speeds. Our simple method achieves competitive performance on Argoverse-HD dataset and improves the AP by 4.9% compared to the strong baseline, validating its effectiveness. Our code will be made available at this https URL.

25.Binary Morphological Neural Network ⬇️

In the last ten years, Convolutional Neural Networks (CNNs) have formed the basis of deep-learning architectures for most computer vision tasks. However, they are not necessarily optimal. For example, mathematical morphology is known to be better suited to deal with binary images. In this work, we create a morphological neural network that handles binary inputs and outputs. We propose their construction inspired by CNNs to formulate layers adapted to such images by replacing convolutions with erosions and dilations. We give explainable theoretical results on whether or not the resulting learned networks are indeed morphological operators. We present promising experimental results designed to learn basic binary operators, and we have made our code publicly available online.

26.DR.VIC: Decomposition and Reasoning for Video Individual Counting ⬇️

Pedestrian counting is a fundamental tool for understanding pedestrian patterns and crowd flow analysis. Existing works (e.g., image-level pedestrian counting, crossline crowd counting et al.) either only focus on the image-level counting or are constrained to the manual annotation of lines. In this work, we propose to conduct the pedestrian counting from a new perspective - Video Individual Counting (VIC), which counts the total number of individual pedestrians in the given video (a person is only counted once). Instead of relying on the Multiple Object Tracking (MOT) techniques, we propose to solve the problem by decomposing all pedestrians into the initial pedestrians who existed in the first frame and the new pedestrians with separate identities in each following frame. Then, an end-to-end Decomposition and Reasoning Network (DRNet) is designed to predict the initial pedestrian count with the density estimation method and reason the new pedestrian's count of each frame with the differentiable optimal transport. Extensive experiments are conducted on two datasets with congested pedestrians and diverse scenes, demonstrating the effectiveness of our method over baselines with great superiority in counting the individual pedestrians. Code: this https URL.

27.Autofocus for Event Cameras ⬇️

Focus control (FC) is crucial for cameras to capture sharp images in challenging real-world scenarios. The autofocus (AF) facilitates the FC by automatically adjusting the focus settings. However, due to the lack of effective AF methods for the recently introduced event cameras, their FC still relies on naive AF like manual focus adjustments, leading to poor adaptation in challenging real-world conditions. In particular, the inherent differences between event and frame data in terms of sensing modality, noise, temporal resolutions, etc., bring many challenges in designing an effective AF method for event cameras. To address these challenges, we develop a novel event-based autofocus framework consisting of an event-specific focus measure called event rate (ER) and a robust search strategy called event-based golden search (EGS). To verify the performance of our method, we have collected an event-based autofocus dataset (EAD) containing well-synchronized frames, events, and focal positions in a wide variety of challenging scenes with severe lighting and motion conditions. The experiments on this dataset and additional real-world scenarios demonstrated the superiority of our method over state-of-the-art approaches in terms of efficiency and accuracy.

28.Self-supervised HDR Imaging from Motion and Exposure Cues ⬇️

Recent High Dynamic Range (HDR) techniques extend the capabilities of current cameras where scenes with a wide range of illumination can not be accurately captured with a single low-dynamic-range (LDR) image. This is generally accomplished by capturing several LDR images with varying exposure values whose information is then incorporated into a merged HDR image. While such approaches work well for static scenes, dynamic scenes pose several challenges, mostly related to the difficulty of finding reliable pixel correspondences. Data-driven approaches tackle the problem by learning an end-to-end mapping with paired LDR-HDR training data, but in practice generating such HDR ground-truth labels for dynamic scenes is time-consuming and requires complex procedures that assume control of certain dynamic elements of the scene (e.g. actor pose) and repeatable lighting conditions (stop-motion capturing). In this work, we propose a novel self-supervised approach for learnable HDR estimation that alleviates the need for HDR ground-truth labels. We propose to leverage the internal statistics of LDR images to create HDR pseudo-labels. We separately exploit static and well-exposed parts of the input images, which in conjunction with synthetic illumination clipping and motion augmentation provide high quality training examples. Experimental results show that the HDR models trained using our proposed self-supervision approach achieve performance competitive with those trained under full supervision, and are to a large extent superior to previous methods that equally do not require any supervision.

29.Domain-Generalized Textured Surface Anomaly Detection ⬇️

Anomaly detection aims to identify abnormal data that deviates from the normal ones, while typically requiring a sufficient amount of normal data to train the model for performing this task. Despite the success of recent anomaly detection methods, performing anomaly detection in an unseen domain remain a challenging task. In this paper, we address the task of domain-generalized textured surface anomaly detection. By observing normal and abnormal surface data across multiple source domains, our model is expected to be generalized to an unseen textured surface of interest, in which only a small number of normal data can be observed during testing. Although with only image-level labels observed in the training data, our patch-based meta-learning model exhibits promising generalization ability: not only can it generalize to unseen image domains, but it can also localize abnormal regions in the query image. Our experiments verify that our model performs favorably against state-of-the-art anomaly detection and domain generalization approaches in various settings.

30.Lane detection with Position Embedding ⬇️

Recently, lane detection has made great progress in autonomous driving. RESA (REcurrent Feature-Shift Aggregator) is based on image segmentation. It presents a novel module to enrich lane feature after preliminary feature extraction with an ordinary CNN. For Tusimple dataset, there is not too complicated scene and lane has more prominent spatial features. On the basis of RESA, we introduce the method of position embedding to enhance the spatial features. The experimental results show that this method has achieved the best accuracy 96.93% on Tusimple dataset.

31.DAN: a Segmentation-free Document Attention Network for Handwritten Document Recognition ⬇️

Unconstrained handwritten document recognition is a challenging computer vision task. It is traditionally handled by a two-step approach combining line segmentation followed by text line recognition. For the first time, we propose an end-to-end segmentation-free architecture for the task of handwritten document recognition: the Document Attention Network. In addition to the text recognition, the model is trained to label text parts using begin and end tags in an XML-like fashion. This model is made up of an FCN encoder for feature extraction and a stack of transformer decoder layers for a recurrent token-by-token prediction process. It takes whole text documents as input and sequentially outputs characters, as well as logical layout tokens. Contrary to the existing segmentation-based approaches, the model is trained without using any segmentation label. We achieve competitive results on the READ dataset at page level, as well as double-page level with a CER of 3.53% and 3.69%, respectively. We also provide results for the RIMES dataset at page level, reaching 4.54% of CER.
We provide all source code and pre-trained model weights at this https URL.

32.Event-Based Dense Reconstruction Pipeline ⬇️

Event cameras are a new type of sensors that are different from traditional cameras. Each pixel is triggered asynchronously by event. The trigger event is the change of the brightness irradiated on the pixel. If the increment or decrement of brightness is higher than a certain threshold, an event is output. Compared with traditional cameras, event cameras have the advantages of high dynamic range and no motion blur. Since events are caused by the apparent motion of intensity edges, the majority of 3D reconstructed maps consist only of scene edges, i.e., semi-dense maps, which is not enough for some applications. In this paper, we propose a pipeline to realize event-based dense reconstruction. First, deep learning is used to reconstruct intensity images from events. And then, structure from motion (SfM) is used to estimate camera intrinsic, extrinsic and sparse point cloud. Finally, multi-view stereo (MVS) is used to complete dense reconstruction.

33.Ev-TTA: Test-Time Adaptation for Event-Based Object Recognition ⬇️

We introduce Ev-TTA, a simple, effective test-time adaptation algorithm for event-based object recognition. While event cameras are proposed to provide measurements of scenes with fast motions or drastic illumination changes, many existing event-based recognition algorithms suffer from performance deterioration under extreme conditions due to significant domain shifts. Ev-TTA mitigates the severe domain gaps by fine-tuning the pre-trained classifiers during the test phase using loss functions inspired by the spatio-temporal characteristics of events. Since the event data is a temporal stream of measurements, our loss function enforces similar predictions for adjacent events to quickly adapt to the changed environment online. Also, we utilize the spatial correlations between two polarities of events to handle noise under extreme illumination, where different polarities of events exhibit distinctive noise distributions. Ev-TTA demonstrates a large amount of performance gain on a wide range of event-based object recognition tasks without extensive additional training. Our formulation can be successfully applied regardless of input representations and further extended into regression tasks. We expect Ev-TTA to provide the key technique to deploy event-based vision algorithms in challenging real-world applications where significant domain shift is inevitable.

34.Scale-Equivalent Distillation for Semi-Supervised Object Detection ⬇️

Recent Semi-Supervised Object Detection (SS-OD) methods are mainly based on self-training, i.e., generating hard pseudo-labels by a teacher model on unlabeled data as supervisory signals. Although they achieved certain success, the limited labeled data in semi-supervised learning scales up the challenges of object detection. We analyze the challenges these methods meet with the empirical experiment results. We find that the massive False Negative samples and inferior localization precision lack consideration. Besides, the large variance of object sizes and class imbalance (i.e., the extreme ratio between background and object) hinder the performance of prior arts. Further, we overcome these challenges by introducing a novel approach, Scale-Equivalent Distillation (SED), which is a simple yet effective end-to-end knowledge distillation framework robust to large object size variance and class imbalance. SED has several appealing benefits compared to the previous works. (1) SED imposes a consistency regularization to handle the large scale variance problem. (2) SED alleviates the noise problem from the False Negative samples and inferior localization precision. (3) A re-weighting strategy can implicitly screen the potential foreground regions of the unlabeled data to reduce the effect of class imbalance. Extensive experiments show that SED consistently outperforms the recent state-of-the-art methods on different datasets with significant margins. For example, it surpasses the supervised counterpart by more than 10 mAP when using 5% and 10% labeled data on MS-COCO.

35.A Method of Data Augmentation to Train a Small Area Fingerprint Recognition Deep Neural Network with a Normal Fingerprint Database ⬇️

Fingerprints are popular among the biometric based systems due to ease of acquisition, uniqueness and availability. Nowadays it is used in smart phone security, digital payment and digital locker. The traditional fingerprint matching methods based on minutiae are mainly applicable for large-area fingerprint and the accuracy rate would reduce significantly when dealing with small-area fingerprint from smart phone. There are many attempts to using deep learning for small-area fingerprint recognition, and there are many successes. But training deep neural network needs a lot of datasets for training. There is no well-known dataset for small-area, so we have to make datasets ourselves. In this paper, we propose a method of data augmentation to train a small-area fingerprint recognition deep neural network with a normal fingerprint database (such as FVC2002) and verify it via tests. The experimental results showed the efficiency of our method.

36.Negative Selection by Clustering for Contrastive Learning in Human Activity Recognition ⬇️

Contrastive learning has been applied to Human Activity Recognition (HAR) based on sensor data owing to its ability to achieve performance comparable to supervised learning with a large amount of unlabeled data and a small amount of labeled data. The pre-training task for contrastive learning is generally instance discrimination, which specifies that each instance belongs to a single class, but this will consider the same class of samples as negative examples. Such a pre-training task is not conducive to human activity recognition tasks, which are mainly classification tasks. To address this problem, we follow SimCLR to propose a new contrastive learning framework that negative selection by clustering in HAR, which is called ClusterCLHAR. Compared with SimCLR, it redefines the negative pairs in the contrastive loss function by using unsupervised clustering methods to generate soft labels that mask other samples of the same cluster to avoid regarding them as negative samples. We evaluate ClusterCLHAR on three benchmark datasets, USC-HAD, MotionSense, and UCI-HAR, using mean F1-score as the evaluation metric. The experiment results show that it outperforms all the state-of-the-art methods applied to HAR in self-supervised learning and semi-supervised learning.

37.Efficient Few-Shot Object Detection via Knowledge Inheritance ⬇️

Few-shot object detection (FSOD), which aims at learning a generic detector that can adapt to unseen tasks with scarce training samples, has witnessed consistent improvement recently. However, most existing methods ignore the efficiency issues, e.g., high computational complexity and slow adaptation speed. Notably, efficiency has become an increasingly important evaluation metric for few-shot techniques due to an emerging trend toward embedded AI. To this end, we present an efficient pretrain-transfer framework (PTF) baseline with no computational increment, which achieves comparable results with previous state-of-the-art (SOTA) methods. Upon this baseline, we devise an initializer named knowledge inheritance (KI) to reliably initialize the novel weights for the box classifier, which effectively facilitates the knowledge transfer process and boosts the adaptation speed. Within the KI initializer, we propose an adaptive length re-scaling (ALR) strategy to alleviate the vector length inconsistency between the predicted novel weights and the pretrained base weights. Finally, our approach not only achieves the SOTA results across three public benchmarks, i.e., PASCAL VOC, COCO and LVIS, but also exhibits high efficiency with 1.8-9.0x faster adaptation speed against the other methods on COCO/LVIS benchmark during few-shot transfer. To our best knowledge, this is the first work to consider the efficiency problem in FSOD. We hope to motivate a trend toward powerful yet efficient few-shot technique development. The codes are publicly available at this https URL.

38.Training-free Transformer Architecture Search ⬇️

Recently, Vision Transformer (ViT) has achieved remarkable success in several computer vision tasks. The progresses are highly relevant to the architecture design, then it is worthwhile to propose Transformer Architecture Search (TAS) to search for better ViTs automatically. However, current TAS methods are time-consuming and existing zero-cost proxies in CNN do not generalize well to the ViT search space according to our experimental observations. In this paper, for the first time, we investigate how to conduct TAS in a training-free manner and devise an effective training-free TAS (TF-TAS) scheme. Firstly, we observe that the properties of multi-head self-attention (MSA) and multi-layer perceptron (MLP) in ViTs are quite different and that the synaptic diversity of MSA affects the performance notably. Secondly, based on the observation, we devise a modular strategy in TF-TAS that evaluates and ranks ViT architectures from two theoretical perspectives: synaptic diversity and synaptic saliency, termed as DSS-indicator. With DSS-indicator, evaluation results are strongly correlated with the test accuracies of ViT models. Experimental results demonstrate that our TF-TAS achieves a competitive performance against the state-of-the-art manually or automatically design ViT architectures, and it promotes the searching efficiency in ViT search space greatly: from about $24$ GPU days to less than $0.5$ GPU days. Moreover, the proposed DSS-indicator outperforms the existing cutting-edge zero-cost approaches (e.g., TE-score and NASWOT).

39.Self-supervised Learning of Adversarial Example: Towards Good Generalizations for Deepfake Detection ⬇️

Recent studies in deepfake detection have yielded promising results when the training and testing face forgeries are from the same dataset. However, the problem remains challenging when one tries to generalize the detector to forgeries created by unseen methods in the training dataset. This work addresses the generalizable deepfake detection from a simple principle: a generalizable representation should be sensitive to diverse types of forgeries. Following this principle, we propose to enrich the "diversity" of forgeries by synthesizing augmented forgeries with a pool of forgery configurations and strengthen the "sensitivity" to the forgeries by enforcing the model to predict the forgery configurations. To effectively explore the large forgery augmentation space, we further propose to use the adversarial training strategy to dynamically synthesize the most challenging forgeries to the current model. Through extensive experiments, we show that the proposed strategies are surprisingly effective (see Figure 1), and they could achieve superior performance than the current state-of-the-art methods. Code is available at \url{this https URL}.

40.Interpretable Prediction of Lung Squamous Cell Carcinoma Recurrence With Self-supervised Learning ⬇️

Lung squamous cell carcinoma (LSCC) has a high recurrence and metastasis rate. Factors influencing recurrence and metastasis are currently unknown and there are no distinct histopathological or morphological features indicating the risks of recurrence and metastasis in LSCC. Our study focuses on the recurrence prediction of LSCC based on H&E-stained histopathological whole-slide images (WSI). Due to the small size of LSCC cohorts in terms of patients with available recurrence information, standard end-to-end learning with various convolutional neural networks for this task tends to overfit. Also, the predictions made by these models are hard to interpret. Histopathology WSIs are typically very large and are therefore processed as a set of smaller tiles. In this work, we propose a novel conditional self-supervised learning (SSL) method to learn representations of WSI at the tile level first, and leverage clustering algorithms to identify the tiles with similar histopathological representations. The resulting representations and clusters from self-supervision are used as features of a survival model for recurrence prediction at the patient level. Using two publicly available datasets from TCGA and CPTAC, we show that our LSCC recurrence prediction survival model outperforms both LSCC pathological stage-based approach and machine learning baselines such as multiple instance learning. The proposed method also enables us to explain the recurrence histopathological risk factors via the derived clusters. This can help pathologists derive new hypotheses regarding morphological features associated with LSCC recurrence.

41.Deep Frequency Filtering for Domain Generalization ⬇️

Improving the generalization capability of Deep Neural Networks (DNNs) is critical for their practical uses, which has been a longstanding challenge. Some theoretical studies have revealed that DNNs have preferences to different frequency components in the learning process and indicated that this may affect the robustness of learned features. In this paper, we propose Deep Frequency Filtering (DFF) for learning domain-generalizable features, which is the first endeavour to explicitly modulate frequency components of different transfer difficulties across domains during training. To achieve this, we perform Fast Fourier Transform (FFT) on feature maps at different layers, then adopt a light-weight module to learn the attention masks from frequency representations after FFT to enhance transferable frequency components while suppressing the components not conductive to generalization. Further, we empirically compare different types of attention for implementing our conceptualized DFF. Extensive experiments demonstrate the effectiveness of the proposed DFF and show that applying DFF on a plain baseline outperforms the state-of-the-art methods on different domain generalization tasks, including close-set classification and open-set retrieval.

42.Biceph-Net: A robust and lightweight framework for the diagnosis of Alzheimer's disease using 2D-MRI scans and deep similarity learning ⬇️

Alzheimer's Disease (AD) is a neurodegenerative disease that is one of the significant causes of death in the elderly population. Many deep learning techniques have been proposed to diagnose AD using Magnetic Resonance Imaging (MRI) scans. Predicting AD using 2D slices extracted from 3D MRI scans is challenging as the inter-slice information gets lost. To this end, we propose a novel and lightweight framework termed 'Biceph-Net' for AD diagnosis using 2D MRI scans that model both the intra-slice and inter-slice information. Biceph-Net has been experimentally shown to perform similar to other Spatio-temporal neural networks while being computationally more efficient. Biceph-Net is also superior in performance compared to vanilla 2D convolutional neural networks (CNN) for AD diagnosis using 2D MRI slices. Biceph-Net also has an inbuilt neighbourhood-based model interpretation feature that can be exploited to understand the classification decision taken by the network. Biceph-Net experimentally achieves a test accuracy of 100% in the classification of Cognitively Normal (CN) vs AD, 98.16% for Mild Cognitive Impairment (MCI) vs AD, and 97.80% for CN vs MCI vs AD.

43.Self-Supervised Robust Scene Flow Estimation via the Alignment of Probability Density Functions ⬇️

In this paper, we present a new self-supervised scene flow estimation approach for a pair of consecutive point clouds. The key idea of our approach is to represent discrete point clouds as continuous probability density functions using Gaussian mixture models. Scene flow estimation is therefore converted into the problem of recovering motion from the alignment of probability density functions, which we achieve using a closed-form expression of the classic Cauchy-Schwarz divergence. Unlike existing nearest-neighbor-based approaches that use hard pairwise correspondences, our proposed approach establishes soft and implicit point correspondences between point clouds and generates more robust and accurate scene flow in the presence of missing correspondences and outliers. Comprehensive experiments show that our method makes noticeable gains over the Chamfer Distance and the Earth Mover's Distance in real-world environments and achieves state-of-the-art performance among self-supervised learning methods on FlyingThings3D and KITTI, even outperforming some supervised methods with ground truth annotations.

44.Learning to Censor by Noisy Sampling ⬇️

Point clouds are an increasingly ubiquitous input modality and the raw signal can be efficiently processed with recent progress in deep learning. This signal may, often inadvertently, capture sensitive information that can leak semantic and geometric properties of the scene which the data owner does not want to share. The goal of this work is to protect sensitive information when learning from point clouds; by censoring the sensitive information before the point cloud is released for downstream tasks. Specifically, we focus on preserving utility for perception tasks while mitigating attribute leakage attacks. The key motivating insight is to leverage the localized saliency of perception tasks on point clouds to provide good privacy-utility trade-offs. We realize this through a mechanism called Censoring by Noisy Sampling (CBNS), which is composed of two modules: i) Invariant Sampler: a differentiable point-cloud sampler which learns to remove points invariant to utility and ii) Noisy Distorter: which learns to distort sampled points to decouple the sensitive information from utility, and mitigate privacy leakage. We validate the effectiveness of CBNS through extensive comparisons with state-of-the-art baselines and sensitivity analyses of key design choices. Results show that CBNS achieves superior privacy-utility trade-offs on multiple datasets.

45.Unifying Motion Deblurring and Frame Interpolation with Events ⬇️

Slow shutter speed and long exposure time of frame-based cameras often cause visual blur and loss of inter-frame information, degenerating the overall quality of captured videos. To this end, we present a unified framework of event-based motion deblurring and frame interpolation for blurry video enhancement, where the extremely low latency of events is leveraged to alleviate motion blur and facilitate intermediate frame prediction. Specifically, the mapping relation between blurry frames and sharp latent images is first predicted by a learnable double integral network, and a fusion network is then proposed to refine the coarse results via utilizing the information from consecutive blurry inputs and the concurrent events. By exploring the mutual constraints among blurry frames, latent images, and event streams, we further propose a self-supervised learning framework to enable network training with real-world blurry videos and events. Extensive experiments demonstrate that our method compares favorably against the state-of-the-art approaches and achieves remarkable performance on both synthetic and real-world datasets.

46.Adaptive Transformers for Robust Few-shot Cross-domain Face Anti-spoofing ⬇️

While recent face anti-spoofing methods perform well under the intra-domain setups, an effective approach needs to account for much larger appearance variations of images acquired in complex scenes with different sensors for robust performance. In this paper, we present adaptive vision transformers (ViT) for robust cross-domain face anti-spoofing. Specifically, we adopt ViT as a backbone to exploit its strength to account for long-range dependencies among pixels. We further introduce the ensemble adapters module and feature-wise transformation layers in the ViT to adapt to different domains for robust performance with a few samples. Experiments on several benchmark datasets show that the proposed models achieve both robust and competitive performance against the state-of-the-art methods.

47.Comprehensive Benchmark Datasets for Amharic Scene Text Detection and Recognition ⬇️

Ethiopic/Amharic script is one of the oldest African writing systems, which serves at least 23 languages (e.g., Amharic, Tigrinya) in East Africa for more than 120 million people. The Amharic writing system, Abugida, has 282 syllables, 15 punctuation marks, and 20 numerals. The Amharic syllabic matrix is derived from 34 base graphemes/consonants by adding up to 12 appropriate diacritics or vocalic markers to the characters. The syllables with a common consonant or vocalic markers are likely to be visually similar and challenge text recognition tasks. In this work, we presented the first comprehensive public datasets named HUST-ART, HUST-AST, ABE, and Tana for Amharic script detection and recognition in the natural scene. We have also conducted extensive experiments to evaluate the performance of the state of art methods in detecting and recognizing Amharic scene text on our datasets. The evaluation results demonstrate the robustness of our datasets for benchmarking and its potential of promoting the development of robust Amharic script detection and recognition algorithms. Consequently, the outcome will benefit people in East Africa, including diplomats from several countries and international communities.

48.Semi-Supervised Hybrid Spine Network for Segmentation of Spine MR Images ⬇️

Automatic segmentation of vertebral bodies (VBs) and intervertebral discs (IVDs) in 3D magnetic resonance (MR) images is vital in diagnosing and treating spinal diseases. However, segmenting the VBs and IVDs simultaneously is not trivial. Moreover, problems exist, including blurry segmentation caused by anisotropy resolution, high computational cost, inter-class similarity and intra-class variability, and data imbalances. We proposed a two-stage algorithm, named semi-supervised hybrid spine network (SSHSNet), to address these problems by achieving accurate simultaneous VB and IVD segmentation. In the first stage, we constructed a 2D semi-supervised DeepLabv3+ by using cross pseudo supervision to obtain intra-slice features and coarse segmentation. In the second stage, a 3D full-resolution patch-based DeepLabv3+ was built. This model can be used to extract inter-slice information and combine the coarse segmentation and intra-slice features provided from the first stage. Moreover, a cross tri-attention module was applied to compensate for the loss of inter-slice and intra-slice information separately generated from 2D and 3D networks, thereby improving feature representation ability and achieving satisfactory segmentation results. The proposed SSHSNet was validated on a publicly available spine MR image dataset, and remarkable segmentation performance was achieved. Moreover, results show that the proposed method has great potential in dealing with the data imbalance problem. Based on previous reports, few studies have incorporated a semi-supervised learning strategy with a cross attention mechanism for spine segmentation. Therefore, the proposed method may provide a useful tool for spine segmentation and aid clinically in spinal disease diagnoses and treatments. Codes are publicly available at: this https URL.

49.Pixel VQ-VAEs for Improved Pixel Art Representation ⬇️

Machine learning has had a great deal of success in image processing. However, the focus of this work has largely been on realistic images, ignoring more niche art styles such as pixel art. Additionally, many traditional machine learning models that focus on groups of pixels do not work well with pixel art, where individual pixels are important. We propose the Pixel VQ-VAE, a specialized VQ-VAE model that learns representations of pixel art. We show that it outperforms other models in both the quality of embeddings as well as performance on downstream tasks.

50.Contrastive Transformer-based Multiple Instance Learning for Weakly Supervised Polyp Frame Detection ⬇️

Current polyp detection methods from colonoscopy videos use exclusively normal (i.e., healthy) training images, which i) ignore the importance of temporal information in consecutive video frames, and ii) lack knowledge about the polyps. Consequently, they often have high detection errors, especially on challenging polyp cases (e.g., small, flat, or partially visible polyps). In this work, we formulate polyp detection as a weakly-supervised anomaly detection task that uses video-level labelled training data to detect frame-level polyps. In particular, we propose a novel convolutional transformer-based multiple instance learning method designed to identify abnormal frames (i.e., frames with polyps) from anomalous videos (i.e., videos containing at least one frame with polyp). In our method, local and global temporal dependencies are seamlessly captured while we simultaneously optimise video and snippet-level anomaly scores. A contrastive snippet mining method is also proposed to enable an effective modelling of the challenging polyp cases. The resulting method achieves a detection accuracy that is substantially better than current state-of-the-art approaches on a new large-scale colonoscopy video dataset introduced in this work.

51.Visual Prompt Tuning ⬇️

The current modus operandi in adapting pre-trained models involves updating all the backbone parameters, ie, full fine-tuning. This paper introduces Visual Prompt Tuning (VPT) as an efficient and effective alternative to full fine-tuning for large-scale Transformer models in vision. Taking inspiration from recent advances in efficiently tuning large language models, VPT introduces only a small amount (less than 1% of model parameters) of trainable parameters in the input space while keeping the model backbone frozen. Via extensive experiments on a wide variety of downstream recognition tasks, we show that VPT achieves significant performance gains compared to other parameter efficient tuning protocols. Most importantly, VPT even outperforms full fine-tuning in many cases across model capacities and training data scales, while reducing per-task storage cost.

52.GOSS: Towards Generalized Open-set Semantic Segmentation ⬇️

In this paper, we present and study a new image segmentation task, called Generalized Open-set Semantic Segmentation (GOSS). Previously, with the well-known open-set semantic segmentation (OSS), the intelligent agent only detects the unknown regions without further processing, limiting their perception of the environment. It stands to reason that a further analysis of the detected unknown pixels would be beneficial. Therefore, we propose GOSS, which unifies the abilities of two well-defined segmentation tasks, OSS and generic segmentation (GS), in a holistic way. Specifically, GOSS classifies pixels as belonging to known classes, and clusters (or groups) of pixels of unknown class are labelled as such. To evaluate this new expanded task, we further propose a metric which balances the pixel classification and clustering aspects. Moreover, we build benchmark tests on top of existing datasets and propose a simple neural architecture as a baseline, which jointly predicts pixel classification and clustering under open-set settings. Our experiments on multiple benchmarks demonstrate the effectiveness of our baseline. We believe our new GOSS task can produce an expressive image understanding for future research. Code will be made available.

53.Fast on-line signature recognition based on VQ with time modeling ⬇️

This paper proposes a multi-section vector quantization approach for on-line signature recognition. We have used the MCYT database, which consists of 330 users and 25 skilled forgeries per person performed by 5 different impostors. This database is larger than those typically used in the literature. Nevertheless, we also provide results from the SVC database.
Our proposed system outperforms the winner of SVC with a reduced computational requirement, which is around 47 times lower than DTW. In addition, our system improves the database storage requirements due to vector compression, and is more privacy-friendly as it is not possible to recover the original signature using the codebooks. Experimental results with MCYT provide a 99.76% identification rate and 2.46% EER (skilled forgeries and individual threshold). Experimental results with SVC are 100% of identification rate and 0% (individual threshold) and 0.31% (general threshold) when using a two-section VQ approach.

54.Deep Portrait Delighting ⬇️

We present a deep neural network for removing undesirable shading features from an unconstrained portrait image, recovering the underlying texture. Our training scheme incorporates three regularization strategies: masked loss, to emphasize high-frequency shading features; soft-shadow loss, which improves sensitivity to subtle changes in lighting; and shading-offset estimation, to supervise separation of shading and texture. Our method demonstrates improved delighting quality and generalization when compared with the state-of-the-art. We further demonstrate how our delighting method can enhance the performance of light-sensitive computer vision tasks such as face relighting and semantic parsing, allowing them to handle extreme lighting conditions.

55.PlaneMVS: 3D Plane Reconstruction from Multi-View Stereo ⬇️

We present a novel framework named PlaneMVS for 3D plane reconstruction from multiple input views with known camera poses. Most previous learning-based plane reconstruction methods reconstruct 3D planes from single images, which highly rely on single-view regression and suffer from depth scale ambiguity. In contrast, we reconstruct 3D planes with a multi-view-stereo (MVS) pipeline that takes advantage of multi-view geometry. We decouple plane reconstruction into a semantic plane detection branch and a plane MVS branch. The semantic plane detection branch is based on a single-view plane detection framework but with differences. The plane MVS branch adopts a set of slanted plane hypotheses to replace conventional depth hypotheses to perform plane sweeping strategy and finally learns pixel-level plane parameters and its planar depth map. We present how the two branches are learned in a balanced way, and propose a soft-pooling loss to associate the outputs of the two branches and make them benefit from each other. Extensive experiments on various indoor datasets show that PlaneMVS significantly outperforms state-of-the-art (SOTA) single-view plane reconstruction methods on both plane detection and 3D geometry metrics. Our method even outperforms a set of SOTA learning-based MVS methods thanks to the learned plane priors. To the best of our knowledge, this is the first work on 3D plane reconstruction within an end-to-end MVS framework.

56.DTFD-MIL: Double-Tier Feature Distillation Multiple Instance Learning for Histopathology Whole Slide Image Classification ⬇️

Multiple instance learning (MIL) has been increasingly used in the classification of histopathology whole slide images (WSIs). However, MIL approaches for this specific classification problem still face unique challenges, particularly those related to small sample cohorts. In these, there are limited number of WSI slides (bags), while the resolution of a single WSI is huge, which leads to a large number of patches (instances) cropped from this slide. To address this issue, we propose to virtually enlarge the number of bags by introducing the concept of pseudo-bags, on which a double-tier MIL framework is built to effectively use the intrinsic features. Besides, we also contribute to deriving the instance probability under the framework of attention-based MIL, and utilize the derivation to help construct and analyze the proposed framework. The proposed method outperforms other latest methods on the CAMELYON-16 by substantially large margins, and is also better in performance on the TCGA lung cancer dataset. The proposed framework is ready to be extended for wider MIL applications. The code is available at: this https URL

57.Self-supervision through Random Segments with Autoregressive Coding (RandSAC) ⬇️

Inspired by the success of self-supervised autoregressive representation learning in natural language (GPT and its variants), and advances in recent visual architecture design with Vision Transformers (ViTs), in this paper, we explore the effects various design choices have on the success of applying such training strategies for visual feature learning. Specifically, we introduce a novel strategy that we call Random Segments with Autoregressive Coding (RandSAC). In RandSAC, we group patch representations (image tokens) into hierarchically arranged segments; within each segment, tokens are predicted in parallel, similar to BERT, while across segment predictions are sequential, similar to GPT. We illustrate that randomized serialization of the segments significantly improves the performance and results in distribution over spatially-long (across-segments) and -short (within-segment) predictions which are effective for feature learning. We illustrate the pertinence of these design choices and explore alternatives on a number of datasets (e.g., CIFAR10, ImageNet). While our pre-training strategy works with vanilla Transformer, we also propose a conceptually simple, but highly effective, addition to the decoder that allows learnable skip-connections to encoder feature layers, which further improves the performance. Our final model, trained on ImageNet, achieves new state-of-the-art linear probing performance 68.3% among comparative predictive self-supervised learning approaches.

58.Learning Geodesic-Aware Local Features from RGB-D Images ⬇️

Most of the existing handcrafted and learning-based local descriptors are still at best approximately invariant to affine image transformations, often disregarding deformable surfaces. In this paper, we take one step further by proposing a new approach to compute descriptors from RGB-D images (where RGB refers to the pixel color brightness and D stands for depth information) that are invariant to isometric non-rigid deformations, as well as to scale changes and rotation. Our proposed description strategies are grounded on the key idea of learning feature representations on undistorted local image patches using surface geodesics. We design two complementary local descriptors strategies to compute geodesic-aware features efficiently: one efficient binary descriptor based on handcrafted binary tests (named GeoBit), and one learning-based descriptor (GeoPatch) with convolutional neural networks (CNNs) to compute features. In different experiments using real and publicly available RGB-D data benchmarks, they consistently outperforms state-of-the-art handcrafted and learning-based image and RGB-D descriptors in matching scores, as well as in object retrieval and non-rigid surface tracking experiments, with comparable processing times. We also provide to the community a new dataset with accurate matching annotations of RGB-D images of different objects (shirts, cloths, paintings, bags), subjected to strong non-rigid deformations, for evaluation benchmark of deformable surface correspondence algorithms.

59.Joint Feature Learning and Relation Modeling for Tracking: A One-Stream Framework ⬇️

The current popular two-stream, two-stage tracking framework extracts the template and the search region features separately and then performs relation modeling, thus the extracted features lack the awareness of the target and have limited target-background discriminability. To tackle the above issue, we propose a novel one-stream tracking (OSTrack) framework that unifies feature learning and relation modeling by bridging the template-search image pairs with bidirectional information flows. In this way, discriminative target-oriented features can be dynamically extracted by mutual guidance. Since no extra heavy relation modeling module is needed and the implementation is highly parallelized, the proposed tracker runs at a fast speed. To further improve the inference efficiency, an in-network candidate early elimination module is proposed based on the strong similarity prior calculated in the one-stream framework. As a unified framework, OSTrack achieves state-of-the-art performance on multiple benchmarks, in particular, it shows impressive results on the one-shot tracking benchmark GOT-10k, i.e., achieving 73.7% AO, improving the existing best result (SwinTrack) by 4.3%. Besides, our method maintains a good performance-speed trade-off and shows faster convergence. The code and models will be available at this https URL.

60.Learning Patch-to-Cluster Attention in Vision Transformer ⬇️

The Vision Transformer (ViT) model is built on the assumption of treating image patches as "visual tokens" and learning patch-to-patch attention. The patch embedding based tokenizer is a workaround in practice and has a semantic gap with respect to its counterpart, the textual tokenizer. The patch-to-patch attention suffers from the quadratic complexity issue, and also makes it non-trivial to explain learned ViT models. To address these issues in ViT models, this paper proposes to learn patch-to-cluster attention (PaCa) based ViT models. Queries in our PaCaViT are based on patches, while keys and values are based on clustering (with a predefined small number of clusters). The clusters are learned end-to-end, leading to better tokenizers and realizing joint clustering-for-attention and attention-for-clustering when deployed in ViT models. The quadratic complexity is relaxed to linear complexity. Also, directly visualizing the learned clusters can reveal how a trained ViT model learns to perform a task (e.g., object detection). In experiments, the proposed PaCa-ViT is tested on CIFAR-100 and ImageNet-1000 image classification, and MS-COCO object detection and instance segmentation. Compared with prior arts, it obtains better performance in classification and comparable performance in detection and segmentation. It is significantly more efficient in COCO due to the linear complexity. The learned clusters are also semantically meaningful and shed light on designing more discriminative yet interpretable ViT models.

61.CM-GAN: Image Inpainting with Cascaded Modulation GAN and Object-Aware Training ⬇️

Recent image inpainting methods have made great progress but often struggle to generate plausible image structures when dealing with large holes in complex images. This is partially due to the lack of effective network structures that can capture both the long-range dependency and high-level semantics of an image. To address these problems, we propose cascaded modulation GAN (CM-GAN), a new network design consisting of an encoder with Fourier convolution blocks that extract multi-scale feature representations from the input image with holes and a StyleGAN-like decoder with a novel cascaded global-spatial modulation block at each scale level. In each decoder block, global modulation is first applied to perform coarse semantic-aware structure synthesis, then spatial modulation is applied on the output of global modulation to further adjust the feature map in a spatially adaptive fashion. In addition, we design an object-aware training scheme to prevent the network from hallucinating new objects inside holes, fulfilling the needs of object removal tasks in real-world scenarios. Extensive experiments are conducted to show that our method significantly outperforms existing methods in both quantitative and qualitative evaluation.

62.Improving the Fairness of Chest X-ray Classifiers ⬇️

Deep learning models have reached or surpassed human-level performance in the field of medical imaging, especially in disease diagnosis using chest x-rays. However, prior work has found that such classifiers can exhibit biases in the form of gaps in predictive performance across protected groups. In this paper, we question whether striving to achieve zero disparities in predictive performance (i.e. group fairness) is the appropriate fairness definition in the clinical setting, over minimax fairness, which focuses on maximizing the performance of the worst-case group. We benchmark the performance of nine methods in improving classifier fairness across these two definitions. We find, consistent with prior work on non-clinical data, that methods which strive to achieve better worst-group performance do not outperform simple data balancing. We also find that methods which achieve group fairness do so by worsening performance for all groups. In light of these results, we discuss the utility of fairness definitions in the clinical setting, advocating for an investigation of the bias-inducing mechanisms in the underlying data generating process whenever possible.

63.R3M: A Universal Visual Representation for Robot Manipulation ⬇️

We study how visual representations pre-trained on diverse human video data can enable data-efficient learning of downstream robotic manipulation tasks. Concretely, we pre-train a visual representation using the Ego4D human video dataset using a combination of time-contrastive learning, video-language alignment, and an L1 penalty to encourage sparse and compact representations. The resulting representation, R3M, can be used as a frozen perception module for downstream policy learning. Across a suite of 12 simulated robot manipulation tasks, we find that R3M improves task success by over 20% compared to training from scratch and by over 10% compared to state-of-the-art visual representations like CLIP and MoCo. Furthermore, R3M enables a Franka Emika Panda arm to learn a range of manipulation tasks in a real, cluttered apartment given just 20 demonstrations. Code and pre-trained models are available at this https URL.

64.GriTS: Grid table similarity metric for table structure recognition ⬇️

In this paper, we propose a new class of evaluation metric for table structure recognition, grid table similarity (GriTS). Unlike prior metrics, GriTS evaluates the correctness of a predicted table directly in its natural form as a matrix. To create a similarity measure between matrices, we generalize the two-dimensional largest common substructure (2D-LCS) problem, which is NP-hard, to the 2D most similar substructures (2D-MSS) problem and propose a polynomial-time heuristic for solving it. We validate empirically using the PubTables-1M dataset that comparison between matrices exhibits more desirable behavior than alternatives for table structure recognition evaluation. GriTS also unifies all three subtasks of cell topology recognition, cell location recognition, and cell content recognition within the same framework, which simplifies the evaluation and enables more meaningful comparisons across different types of structure recognition approaches. Code will be released at this https URL.

65.U-Boost NAS: Utilization-Boosted Differentiable Neural Architecture Search ⬇️

Optimizing resource utilization in target platforms is key to achieving high performance during DNN inference. While optimizations have been proposed for inference latency, memory footprint, and energy consumption, prior hardware-aware neural architecture search (NAS) methods have omitted resource utilization, preventing DNNs to take full advantage of the target inference platforms. Modeling resource utilization efficiently and accurately is challenging, especially for widely-used array-based inference accelerators such as Google TPU. In this work, we propose a novel hardware-aware NAS framework that does not only optimize for task accuracy and inference latency, but also for resource utilization. We also propose and validate a new computational model for resource utilization in inference accelerators. By using the proposed NAS framework and the proposed resource utilization model, we achieve 2.8 - 4x speedup for DNN inference compared to prior hardware-aware NAS methods while attaining similar or improved accuracy in image classification on CIFAR-10 and Imagenet-100 datasets.

66.Cell segmentation from telecentric bright-field transmitted light microscopic images using a Residual Attention U-Net: a case study on HeLa line ⬇️

Living cell segmentation from bright-field light microscopic images is challenging due to the image complexity and temporal changes in the living cells. Recently developed deep learning (DL)-based methods became popular in medical and microscopic image segmentation tasks due to their success and promising outcomes. The main objective of this paper is to develop a deep learning, UNet-based method to segment the living cells of the HeLa line in bright-field transmitted light microscopy. To find the most suitable architecture for our datasets, we have proposed a residual attention U-Net and compared it with an attention and a simple U-Net architecture. The attention mechanism highlights the remarkable features and suppresses activations in the irrelevant image regions. The residual mechanism overcomes with vanishing gradient problem. The Mean-IoU score for our datasets reaches 0.9505, 0.9524, and 0.9530 for the simple, attention, and residual attention U-Net, respectively. We achieved the most accurate semantic segmentation results in the Mean-IoU and Dice metrics by applying the residual and attention mechanisms together. The watershed method applied to this best - Residual Attention - semantic segmentation result gave the segmentation with the specific information for each cell.

67.A Multi-Characteristic Learning Method with Micro-Doppler Signatures for Pedestrian Identification ⬇️

The identification of pedestrians using radar micro-Doppler signatures has become a hot topic in recent years. In this paper, we propose a multi-characteristic learning (MCL) model with clusters to jointly learn discrepant pedestrian micro-Doppler signatures and fuse the knowledge learned from each cluster into final decisions. Time-Doppler spectrogram (TDS) and signal statistical features extracted from FMCW radar, as two categories of micro-Doppler signatures, are used in MCL to learn the micro-motion information inside pedestrians' free walking patterns. The experimental results show that our model achieves a higher accuracy rate and is more stable for pedestrian identification than other studies, which make our model more practical.

68.Physics-Driven Deep Learning for Computational Magnetic Resonance Imaging ⬇️

Physics-driven deep learning methods have emerged as a powerful tool for computational magnetic resonance imaging (MRI) problems, pushing reconstruction performance to new limits. This article provides an overview of the recent developments in incorporating physics information into learning-based MRI reconstruction. We consider inverse problems with both linear and non-linear forward models for computational MRI, and review the classical approaches for solving these. We then focus on physics-driven deep learning approaches, covering physics-driven loss functions, plug-and-play methods, generative models, and unrolled networks. We highlight domain-specific challenges such as real- and complex-valued building blocks of neural networks, and translational applications in MRI with linear and non-linear forward models. Finally, we discuss common issues and open challenges, and draw connections to the importance of physics-driven learning when combined with other downstream tasks in the medical imaging pipeline.

69.Lymphocyte Classification in Hyperspectral Images of Ovarian Cancer Tissue Biopsy Samples ⬇️

Current methods for diagnosing the progression of multiple types of cancer within patients rely on interpreting stained needle biopsies. This process is time-consuming and susceptible to error throughout the paraffinization, Hematoxylin and Eosin (H&E) staining, deparaffinization, and annotation stages. Fourier Transform Infrared (FTIR) imaging has been shown to be a promising alternative to staining for appropriately annotating biopsy cores without the need for deparaffinization or H&E staining with the use of Fourier Transform Infrared (FTIR) images when combined with machine learning to interpret the dense spectral information. We present a machine learning pipeline to segment white blood cell (lymphocyte) pixels in hyperspectral images of biopsy cores. These cells are clinically important for diagnosis, but some prior work has struggled to incorporate them due to difficulty obtaining precise pixel labels. Evaluated methods include Support Vector Machine (SVM), Gaussian Naive Bayes, and Multilayer Perceptron (MLP), as well as analyzing the comparatively modern convolutional neural network (CNN).

70.FxP-QNet: A Post-Training Quantizer for the Design of Mixed Low-Precision DNNs with Dynamic Fixed-Point Representation ⬇️

Deep neural networks (DNNs) have demonstrated their effectiveness in a wide range of computer vision tasks, with the state-of-the-art results obtained through complex and deep structures that require intensive computation and memory. Now-a-days, efficient model inference is crucial for consumer applications on resource-constrained platforms. As a result, there is much interest in the research and development of dedicated deep learning (DL) hardware to improve the throughput and energy efficiency of DNNs. Low-precision representation of DNN data-structures through quantization would bring great benefits to specialized DL hardware. However, the rigorous quantization leads to a severe accuracy drop. As such, quantization opens a large hyper-parameter space at bit-precision levels, the exploration of which is a major challenge. In this paper, we propose a novel framework referred to as the Fixed-Point Quantizer of deep neural Networks (FxP-QNet) that flexibly designs a mixed low-precision DNN for integer-arithmetic-only deployment. Specifically, the FxP-QNet gradually adapts the quantization level for each data-structure of each layer based on the trade-off between the network accuracy and the low-precision requirements. Additionally, it employs post-training self-distillation and network prediction error statistics to optimize the quantization of floating-point values into fixed-point numbers. Examining FxP-QNet on state-of-the-art architectures and the benchmark ImageNet dataset, we empirically demonstrate the effectiveness of FxP-QNet in achieving the accuracy-compression trade-off without the need for training. The results show that FxP-QNet-quantized AlexNet, VGG-16, and ResNet-18 reduce the overall memory requirements of their full-precision counterparts by 7.16x, 10.36x, and 6.44x with less than 0.95%, 0.95%, and 1.99% accuracy drop, respectively.

71.WayFAST: Traversability Predictive Navigation for Field Robots ⬇️

We present a self-supervised approach for learning to predict traversable paths for wheeled mobile robots that require good traction to navigate. Our algorithm, termed WayFAST (Waypoint Free Autonomous Systems for Traversability), uses RGB and depth data, along with navigation experience, to autonomously generate traversable paths in outdoor unstructured environments. Our key inspiration is that traction can be estimated for rolling robots using kinodynamic models. Using traction estimates provided by an online receding horizon estimator, we are able to train a traversability prediction neural network in a self-supervised manner, without requiring heuristics utilized by previous methods. We demonstrate the effectiveness of WayFAST through extensive field testing in varying environments, ranging from sandy dry beaches to forest canopies and snow covered grass fields. Our results clearly demonstrate that WayFAST can learn to avoid geometric obstacles as well as untraversable terrain, such as snow, which would be difficult to avoid with sensors that provide only geometric data, such as LiDAR. Furthermore, we show that our training pipeline based on online traction estimates is more data-efficient than other heuristic-based methods.

72.Generative Modeling Helps Weak Supervision (and Vice Versa) ⬇️

Many promising applications of supervised machine learning face hurdles in the acquisition of labeled data in sufficient quantity and quality, creating an expensive bottleneck. To overcome such limitations, techniques that do not depend on ground truth labels have been developed, including weak supervision and generative modeling. While these techniques would seem to be usable in concert, improving one another, how to build an interface between them is not well-understood. In this work, we propose a model fusing weak supervision and generative adversarial networks. It captures discrete variables in the data alongside the weak supervision derived label estimate. Their alignment allows for better modeling of sample-dependent accuracies of the weak supervision sources, improving the unobserved ground truth estimate. It is the first approach to enable data augmentation through weakly supervised synthetic images and pseudolabels. Additionally, its learned discrete variables can be inspected qualitatively. The model outperforms baseline weak supervision label models on a number of multiclass classification datasets, improves the quality of generated images, and further improves end-model performance through data augmentation with synthetic samples.