Deep Learning Approach for Human Action Recognition Using a Time Saliency Map Based on Motion Features Considering Camera Movement and Shot in Video Image Sequences.
Abdorreza Alavi GharahbaghVahid HajihashemiJosé J. M. MachadoJoão Manuel R. S. TavaresPublished in: Inf. (2023)
Keyphrases
- saliency map
- video frames
- video sequences
- video data
- image sequences
- video shots
- video content
- motion analysis
- visual features
- key frames
- motion vectors
- motion patterns
- visual attention
- motion estimation
- video analysis
- video retrieval
- object tracking
- optical flow
- motion model
- moving objects
- moving camera
- human actions
- human motion
- space time
- natural images
- flow field
- human visual system
- motion segmentation
- low level features
- object segmentation
- low level
- three dimensional
- face recognition
- action recognition