Improving the matching of graphs generated from shapes by the use of procrustes distances into a clique-based MAP formulation.
Gerard SanromaFrancesc SerratosaRené AlquézarPublished in: ICPR (2008)
Keyphrases
- graph matching
- procrustes analysis
- shape analysis
- shape matching
- hausdorff distance
- matching algorithm
- shape context
- maximum clique
- association graph
- shape recognition
- maximum weight
- independent set
- subgraph isomorphism
- shape descriptors
- shape decomposition
- pattern matching
- partial matching
- maximum a posteriori probability
- graph embedding
- object recognition
- distance measure
- fourier descriptors
- matching process
- distance function
- shape prior
- maximum a posteriori
- map matching
- attributed relational graph
- shape similarity
- dissimilarity measure
- maximum cardinality
- euclidean distance
- graph mining
- quasi cliques
- matching scheme
- similarity measure
- shape representation
- graph theory
- image matching
- keypoints
- feature points
- optimal solution