Convolutional neural networks on irregular domains through approximate translations on inferred graphs.
Bastien PasdeloupVincent GriponJean-Charles VialatteDominique PastorPublished in: CoRR (2017)
Keyphrases
- convolutional neural networks
- application domains
- real world
- graph theory
- machine translation
- graph theoretic
- convolutional network
- transfer learning
- cross domain
- graph partitioning
- weighted graph
- subgraph isomorphism
- arbitrarily shaped
- graph structures
- search problems
- graph representation
- graph structure
- graph matching
- directed graph
- object detection
- facial expressions