Multi-Task Learning of Generalizable Representations for Video Action Recognition.
Zhiyu YaoYunbo WangMingsheng LongJianmin WangPhilip S. YuJiaguang SunPublished in: ICME (2020)
Keyphrases
- action recognition
- multi task learning
- human actions
- action classification
- multi task
- action detection
- recognizing human actions
- recognition of human actions
- learning tasks
- gaussian processes
- human activities
- activity recognition
- max margin
- bag of words
- space time interest points
- human detection
- body parts
- theoretical analysis
- video sequences
- computer vision
- learning problems
- transfer learning
- high order
- spatio temporal
- learning algorithm
- human pose
- inductive learning
- video data
- space time
- multi class
- metric learning
- knn
- feature vectors
- learning models
- maximum likelihood
- pairwise
- reinforcement learning
- pose estimation