Loading...
Please wait, while we are loading the content...
Similar Documents
A comparison of two energy minimization methods for graph matching.
| Content Provider | CiteSeerX |
|---|---|
| Author | Abdulrahim, Mohammad A. Misra, Manavendra |
| Abstract | . Two approaches for solving the graph matching problem are discussed. Among the nonlinear optimization approaches that are inspired by biology, the Dynamic Link Architecture (DLA) is discussed. The DLA presents a learning style where significant relations in the input patterns are recognized and expressed by the unsupervised selforganization of the components of the DLA. To evaluate the efficiency of the DLA approach, the Graduated Assignment (GA) algorithm is presented and discussed. The paper concludes with a presentation of the attractiveness of the DLA in solving the graph matching problem for purposes of invariant pattern recognition. 1 Introduction Graph matching is used in a broad variety of fields for pattern recognition and classification purposes. One such field is medicine, where one may need to identify malignant cells in an image. To help biologists uncover the semantics of proteins, researchers have presented computational approaches to discovering these semantics in mo... |
| File Format | |
| Access Restriction | Open |
| Subject Keyword | Graph Matching Energy Minimization Method Graph Matching Problem Classification Purpose Unsupervised Selforganization Pattern Recognition Input Pattern Graduated Assignment Broad Variety Significant Relation Introduction Graph Matching Malignant Cell Learning Style Computational Approach Invariant Pattern Recognition Dynamic Link Architecture Nonlinear Optimization Approach Dla Approach |
| Content Type | Text |
