Dynamically encoded actions based on spacetime saliency.
Christoph FeichtenhoferAxel PinzRichard P. WildesPublished in: CVPR (2015)
Keyphrases
- space time
- human actions
- spatial and temporal
- video sequences
- spatio temporal
- dynamic scenes
- changing environment
- motion patterns
- salient points
- structured light
- physically plausible
- saliency map
- image sequences
- reasoning about actions
- visual attention
- human activities
- prior knowledge
- salient regions
- multi view
- salient object detection
- image features