Parameters Optimization of Synergetic Recognition Approach

摘要

Synergetic pattern recognition is a novel and effective pattern recognition method, and has some advantages in image recognition. Researches have shown that attention parameters λ and parameters B， C directly influence on the recognition results, but there is no general research theory to control these parameters in the recognition process. We abstractly analyze these parameters in this paper, and purpose a novel parameters optimization method based on simulated annealing algorithm. SA algorithm has good optimization performance and is used to search the global optimized solution of these parameters. Theoretic analysis and experimental results both show that the proposed parameters optimization method is effective, which can fully improve the performance of synergetic recognition approach, and the algorithm realization is simple and fast.