DDLSTM: Dual-Domain LSTM for Cross-Dataset Action Recognition.
Toby PerrettDima DamenPublished in: CVPR (2019)
Keyphrases
- action recognition
- human actions
- bag of words
- space time interest points
- spatial temporal
- ucf sports
- video dataset
- body parts
- action classification
- computer vision
- recognizing human actions
- human detection
- activity recognition
- depth sensors
- action recognition in videos
- action detection
- human pose
- action primitives
- mid level
- bag of features
- human activities
- recognizing actions
- view invariant
- recognition of human actions
- visual features
- video sequences
- moving objects