Cross-View Action Recognition Using View-Invariant Pose Feature Learned from Synthetic Data with Domain Adaptation.
Yu-Huan YangAn-Sheng LiuYu-Hung LiuTso-Hsin YehZi-Jun LiLi-Chen FuPublished in: ACCV (2) (2018)
Keyphrases
- synthetic data
- domain adaptation
- view invariant
- action recognition
- multi view
- single view
- human actions
- cross domain
- co training
- real world
- semi supervised
- multiple sources
- human motion
- labeled data
- viewpoint
- d objects
- data sets
- target domain
- feature representation
- test data
- transfer learning
- pose estimation
- feature vectors
- reinforcement learning
- image features
- spatio temporal
- computer vision
- unlabeled data
- feature extraction