1. Open-World Object Detection [OW-DETR]
2. Weakly-Supervised Learning
Temporal Action Localization [3C-Net] [D2-Net]
3. Zero- and Few-Shot Learning
Few-Shot Action Recognition [STRM]
Zero-Shot Image Recognition [TF-VAEGAN] [BiAM]
Generalized Zero-Shot Action Recognition [GZSL-OD]
4. Action/Gesture Recognition in videos
5. First-person (egocentric) vision
6. Human pose estimation
7. Patch Matching [PS-Dataset]