Semi-supervised Learning Combining 2DCNNs and Video Compression for Action Recognition.
Hayato TeraoWataru NoguchiHiroyuki IizukaMasahito YamamotoPublished in: ICVISP (2020)
Keyphrases
- action recognition
- semi supervised learning
- video compression
- motion compensation
- semi supervised
- unlabeled data
- video coding
- motion estimation
- human actions
- unsupervised learning
- motion compensated
- labeled data
- video data
- bag of words
- supervised learning
- machine learning
- training data
- compression ratio
- computer vision
- active learning
- motion vectors
- probabilistic model
- video sequences
- multimedia
- image sequences
- transfer learning
- pairwise
- image classification
- object detection