Just One Moment: Structural Vulnerability of Deep Action Recognition against One Frame Attack.
Jaehui HwangJun-Hyuk KimJun-Ho ChoiJong-Seok LeePublished in: ICCV (2021)
Keyphrases
- action recognition
- bag of words
- attack graphs
- human actions
- computer vision
- activity recognition
- human detection
- spatial temporal
- body parts
- action classification
- recognition of human actions
- attack graph
- recognizing human actions
- action primitives
- human activities
- human pose
- frame rate
- bag of features
- depth sensors
- independent subspace analysis
- action detection
- key frames
- video frames
- view invariant
- video dataset
- static images
- mid level
- object detection
- image features
- action recognition in videos
- optical flow