FCSS: Fully Convolutional Self-Similarity for Dense Semantic Correspondence.
Seungryong KimDongbo MinBumsub HamSangryul JeonStephen LinKwanghoon SohnPublished in: CoRR (2017)
Keyphrases
- point to point correspondences
- dense correspondence
- stereo correspondence
- wide baseline stereo
- semantic network
- semantic knowledge
- semantic search
- natural images
- sparse coding
- search engine
- semantic web
- domain specific
- natural language
- point correspondences
- semantic annotation
- network architecture
- semantic features
- image matching
- domain ontology
- deep learning
- semantically equivalent