ADCI-Net: an adaptive discriminative clip identification strategy for fast video action recognition.
Zizhao GuoXiujuan ZhengXinyi ChunSancong YingPublished in: J. Electronic Imaging (2021)
Keyphrases
- action recognition
- action classification
- human actions
- mid level
- action detection
- video dataset
- spatial temporal
- video clips
- bag of words
- recognizing actions
- recognition of human actions
- motion features
- human detection
- recognizing human actions
- deformable part models
- static images
- computer vision
- activity recognition
- space time interest points
- human activities
- max margin
- body parts
- mid level features
- video streams
- motion history images
- human pose
- video data
- view invariant
- key frames
- semi supervised
- video content
- video sequences
- depth sensors
- view invariant action recognition
- action recognition in videos
- bag of features
- video analysis
- video retrieval
- space time
- keywords