SSSC-AM: A Unified Framework for Video Co-Segmentation by Structured Sparse Subspace Clustering with Appearance and Motion Features.
Junlin YaoFrank NielsenPublished in: CoRR (2016)
Keyphrases
- motion features
- motion cues
- subspace clustering
- bounding box
- video data
- high dimensional
- motion analysis
- activity recognition
- high dimensional data
- action recognition
- optical flow
- video shots
- visual features
- motion patterns
- clustering algorithm
- human actions
- color features
- clustering method
- motion vectors
- sparse representation
- motion model
- depth images
- multimedia
- image sequences
- segmentation algorithm
- video retrieval
- flow field
- object segmentation
- computer vision
- video frames
- motion estimation
- shape features
- video clips
- motion segmentation
- video content
- pose estimation
- dimensionality reduction
- segmentation method
- spatio temporal
- shape prior
- data sets