FCSS: Fully Convolutional Self-Similarity for Dense Semantic Correspondence.
Seungryong KimDongbo MinBumsub HamStephen LinKwanghoon SohnPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2019)
Keyphrases
- stereo correspondence
- dense correspondence
- point to point correspondences
- semantic web
- semantic knowledge
- semantic information
- semantic search
- sparse coding
- wide baseline stereo
- deep learning
- semantic network
- low level features
- domain specific
- learning algorithm
- data sets
- semantic annotation
- point correspondences
- semantic analysis
- image matching
- natural images
- semantic representation
- semi supervised
- data sources
- optical flow