Hybrid convolutional neural networks and optical flow for video visual attention prediction.
Meijun SunZiqi ZhouDong ZhangZheng WangPublished in: Multim. Tools Appl. (2018)
Keyphrases
- visual attention
- convolutional neural networks
- optical flow
- visual saliency
- saliency map
- eye tracking data
- eye movements
- eye tracking
- vision system
- visual motion
- focus of attention
- visual attention model
- visual search
- moving camera
- multimedia
- visual input
- video data
- video content
- higher level
- video streams
- video sequences
- salient regions
- convolutional network
- attention mechanism
- image sequences
- motion segmentation
- visual scene
- motion estimation
- object based visual attention
- video summarization
- spatial and temporal
- camera motion
- flow field
- key frames
- event detection
- real time
- motion model
- video coding
- multiresolution
- visual analysis
- natural images
- moving objects
- motion field
- multiscale