Memetic Algorithm Based On Content-based Image Retrieval

Content-Based Image Retrieval (CBIR) is a kind of image analysis technology, which forms the image feature vector by extracting the image information of color, texture, shape and spatial relationship, then image matching, Ultimately realizing Image Retrieval. At present, there are the problems of low efficiency and query accuracy in the CBIR technology. Therefore, intelligent optimization algorithms are applied to CBIR, which more and more scholars are concerned about.In this paper, we will introduce intelligent optimization algorithms with a capacity of a good optimization, for the problems of CBIR technology——Memetic Algorithm (MA). In this paper, MA algorithm use Tabu Search algorithm (TS) as a local search strategy, And use internationally accepted test database for algorithm--TSPLIB to test the performance of the MA algorithm and GA algorithm. Simulation results of two Groups of the test data show that, the solution of MA algorithm is better, and the error is smaller, and demonstrating excellent optimization ability of MA algorithm.Based on the above study, proposing a method of Memetic Algorithm based on Content-Based Image Retrieval. The optimization ability of MA is used to search the best matching image with characteristics in the image library, realizing the process of image retrieval. Simulation results show that, compared with genetic algorithm which is Mature applied to CBIR technologies, this method can effectively improve the precision and recall of image retrieval system. In this paper, it is a success that applying MA to CBIR. There is great significance to the development of CBIR techniques and MA algorithm Applications.