This repository serves as a curated list of research papers, datasets, and benchmarks related to Human-Object Interaction (HOI) learning. It aims to provide a comprehensive overview for researchers and practitioners in computer vision and robotics focusing on understanding and modeling human actions involving objects.

How It Works

The list categorizes HOI research into several key areas: datasets (image, video, 3D), surveys, methods for HOI generation, HOI recognition (image-based), zero-shot/unseen HOI learning, HOI for robotics, HOI detection (instance-based), and specific learning paradigms like few-shot and weakly-supervised HOI. It meticulously lists papers with their publication venues, and often links to their code, project pages, or arXiv preprints.

Quick Start & Requirements

This repository is a reference list and does not have direct installation or execution requirements. It points to external code repositories which may have their own setup instructions.

Highlighted Details

• Extensive coverage of HOI datasets, including BRIGHT, CAL, SynHOI, HICO-DET, V-COCO, and many more.
• Detailed performance benchmarks for various HOI detection methods across different datasets and settings (e.g., HICO-DET, V-COCO).
• Categorization of methods by learning paradigm: zero-shot, few-shot, weakly-supervised, and instance-based detection.
• Inclusion of emerging areas like 3D HOI and HOI for robotics.

Maintenance & Community

The repository is maintained by DirtyHarryLYL. The README encourages contributions for errors or missing publications.

Licensing & Compatibility

The repository itself is a list of links and does not have a specific license. The linked code repositories will have their own licenses.

Limitations & Caveats

The list is a snapshot of research and may not be exhaustive or perfectly up-to-date with the very latest publications. Some links might become outdated over time.