Order-Dependent Convergent Search for Efficient Search of Feature Patterns

Abstract

A novel algorithm, order-dependent convergent search, for efficient search of interest patterns is proposed. Order or sequence in searching plural patterns is analyzed for designing an efficient algorithm in computation cost, and then some schema for reducing the searching area are proposed. For the realization of experimental programs, a matching algorithm of robustness against illumination fluctuation is utilized. For testing the performance of the proposed algorithm, the experimental evaluation using both of synthesized data and real images of industrial products recognized the cost reduction in their registration.