DeepVS: A Deep Learning Based Video Saliency Prediction Approach.
Lai JiangMai XuTie LiuMinglang QiaoZulin WangPublished in: ECCV (14) (2018)
Keyphrases
- deep learning
- unsupervised learning
- unsupervised feature learning
- video sequences
- video data
- machine learning
- deep architectures
- weakly supervised
- multiscale
- visual attention
- active learning
- model selection
- maximum likelihood
- online learning
- video frames
- key frames
- image classification
- mental models
- co occurrence
- pattern recognition
- reinforcement learning