Incremental action recognition and generalizing motion generation based on goal-directed features.
Kathrin GräveSven BehnkePublished in: IROS (2012)
Keyphrases
- action recognition
- goal directed
- static images
- motion features
- human movements
- recognizing human actions
- recognition of human actions
- mid level
- bag of features
- human actions
- action primitives
- bag of words
- computer vision
- action classification
- view invariant
- activity recognition
- human detection
- motion history images
- action recognition in videos
- image sequences
- body parts
- human motion
- image features
- recognizing actions
- motion analysis
- optical flow
- feature vectors
- motion capture data
- action detection
- autonomous robots
- human pose
- real time
- human activities
- motion estimation
- feature space
- feature extraction
- machine learning