DirecFormer: A Directed Attention in Transformer Approach to Robust Action Recognition.
Thanh-Dat TruongQuoc-Huy BuiChi Nhan DuongHan-Seok SeoSon Lam PhungXin LiKhoa LuuPublished in: CVPR (2022)
Keyphrases
- action recognition
- human actions
- view invariant
- bag of words
- human detection
- activity recognition
- action classification
- computer vision
- spatial temporal
- body parts
- human pose
- static images
- recognition of human actions
- recognizing actions
- depth sensors
- action detection
- recognizing human actions
- bag of features
- independent subspace analysis
- view invariant action recognition
- human activities
- viewpoint
- depth cameras
- video dataset
- text classification