DeepFTSG: Multi-stream Asymmetric USE-Net Trellis Encoders with Shared Decoder Feature Fusion Architecture for Video Motion Segmentation.
Gani RahmonKannappan PalaniappanImad Eddine ToubalFiliz BunyakRaghuveer RaoGuna SeetharamanPublished in: Int. J. Comput. Vis. (2024)
Keyphrases
- motion segmentation
- feature fusion
- multi stream
- motion estimation
- low complexity
- video sequences
- image sequences
- moving objects
- motion model
- audio visual
- missing data
- hidden markov models
- feature extraction
- model selection
- motion field
- multiple features
- video content
- multimedia
- spectral clustering
- video data
- optical flow
- space time
- video frames
- video streams
- motion vectors
- image processing
- motion compensation
- feature vectors
- visual data
- computational complexity
- multiscale
- video coding
- image data
- high dimensional
- similarity measure