This repository serves as an extensive, curated collection of research papers, code, and resources related to Temporal Action Detection (TAD), Weakly Supervised Temporal Action Detection (WSTAD), and Temporal Action Proposal Generation (TAPG). It targets researchers and practitioners in computer vision and video understanding, offering a comprehensive overview of the state-of-the-art and historical developments in these fields.

How It Works

The repository is structured by sub-task (detection, proposal generation, weakly supervised) and then chronologically by publication year. It lists papers with links to their code, arXiv pages, or project websites, providing a historical and thematic map of advancements. Key areas covered include transformer-based architectures, self-supervised pre-training, and various supervision strategies from fully supervised to zero-shot and point-level.

Quick Start & Requirements

This repository is a collection of links and does not have a direct installation or execution command. Users will need to follow individual paper links to access specific codebases, which typically require Python environments with deep learning frameworks like PyTorch or TensorFlow, and often specific CUDA versions for GPU acceleration.

Highlighted Details

• Comprehensive coverage of Temporal Action Detection, Weakly Supervised Temporal Action Detection, and Temporal Action Proposal Generation.
• Includes papers from major conferences like CVPR, ICCV, ECCV, AAAI, and TPAMI.
• Features links to code implementations for a vast majority of the listed papers.
• Organizes research by year and sub-task, facilitating historical and thematic exploration.

Maintenance & Community

The repository is maintained by zhenyingfang. Further community interaction details (e.g., Discord, Slack) are not specified in the README.

Licensing & Compatibility

The repository itself is a collection of links and does not have a specific license. Individual code repositories linked within will have their own licenses, which may vary and could include restrictions on commercial use.

Limitations & Caveats

This is a curated list of papers and not a unified framework or library. Users must individually assess and set up the code for each paper they wish to use, which can be time-consuming and require specific dependencies. The rapid pace of research means some links or code might become outdated.