Implementation of ICCV'23 paper "Residual Pattern Learning for Pixel-wise Out-of-Distribution Detection in Semantic Segmentation"

[IJCV 2022] Pytorch codes for Open-set Adversarial Defense with Clean-Adversarial Mutual Learning

PyTorch implementation of our CVPR 2024 paper "Unified Entropy Optimization for Open-Set Test-Time Adaptation"

Python code for detecting and learning new classes of threats present in crops

Source code of PRJ paper "Learning Adversarial Semantic Embeddings for Zero-Shot Recognition in Open Worlds"

Code Implementation of "Unsupervised Recognition of Unknown Objects for Open-World Object Detection"

1st Place Code for FungiCLEF 2023 Competition from UstcAIGroup

A toolbox for one-class classification and open set recognition based on intra-class splitting

Official PyTorch implementation of the paper “Catching Both Gray and Black Swans: Open-set Supervised Anomaly Detection”, open-set anomaly detection, few-shot anomaly detection, semi-supervised anomaly detection.

[ECCV'22 Oral] Pixel-wise Energy-biased Abstention Learning for Anomaly Segmentation on Complex Urban Driving Scenes. Dealing with out-of-distribution detection or open-set recognition in semantic segmentation.

Fooling Machine Learning Models: A Novel Out-of-Distribution Attack through Generative Adversarial Networks

VLG: General Video Recognition with Web Textual Knowledge (https://arxiv.org/abs/2212.01638)

Implementation of CVPR'23 paper "Glocal Energy-based Learning for Few-Shot Open-Set Recognition"

This repository contains the code used to create the results presented in the paper: "From Coarse to Fine-Grained Open-Set Recognition". We investigate the role of label granularity, semantic similarity, and hierarchical representations in open-set recognition (OSR) with an OSR-benchmark based on iNat2021.

Official PyTorch implementation for "SIO: Synthetic In-Distribution Data Benefits Out-of-Distribution Detection"

Official code for paper "OpenCIL: Benchmarking Out-of-Distribution Detection in Class-Incremental Learning"

Source code for baseline obtenience

[CVPR 2022 Oral] Towards Open Set Temporal Action Localization

This is a novel unknown sar target identification method based on feature extraction networks and KLD-RPA joint discrimination. Experiment results form MSTAR dataset shows that our proposed Fea-DA achieves state of the art unknown sar target identification accuracy while maintaining the high recognition accuracy of known target.