FCSS: Fully Convolutional Self-Similarity for Dense Semantic Correspondence.
Seungryong KimDongbo MinBumsub HamSangryul JeonStephen LinKwanghoon SohnPublished in: CVPR (2017)
Keyphrases
- stereo correspondence
- point to point correspondences
- dense correspondence
- high level
- wide baseline stereo
- semantic network
- semantic annotation
- semantic information
- semantic web
- network architecture
- domain specific
- semantic analysis
- semantic search
- computer vision
- sparse coding
- natural images
- context dependent
- semantic relationships
- semantic knowledge
- semantic representation
- motion estimation
- natural language