Learning and inferring human actions with temporal pyramid features based on conditional random fields.
Shih-Yao LinYen-Yu LinChu-Song ChenYi-Ping HungPublished in: ICASSP (2017)
Keyphrases
- conditional random fields
- structured prediction
- sequence labeling
- learning algorithm
- action recognition
- space time
- graphical models
- learning process
- spatio temporal
- human actions
- reinforcement learning
- prior knowledge
- hidden markov models
- supervised learning
- information extraction
- probabilistic model
- parameter learning
- crf model
- information retrieval
- image features
- feature vectors
- generative model
- pairwise
- activity recognition
- three dimensional
- max margin