Performance Evaluation in Color-based Image Retrieval Using Artificial Bee Colony Algorithm

Abstract

This paper focus on the application of artificial bee colony algorithm’s Content Based Image Retrieval (CBIR). The first step is to extract eigenvalues of color moments, color information entropy and color histogram, and then gain a comprehensive color eigenvalues, which is used to represent corresponding image, via weighting the three color eigenvalues after unitization and quantization. The next step to build an image feature library based on the image library. A traditional searching algorithm can be adopted to search the images to get the retrieved results. The third step is to search the images by artificial bee colony algorithm, mainly by global search strategy and local search strategy. In addition, global search strategy and retain elite strategy could accelerate this algorithmic’s astringency, while local search strategy and local area depend on the fitness to improve the accuracy of algorithmic. Finally, the author adopts Recall- Precision Chart to make a comparison between traditional searching algorithm and artificial bee colony algorithm. The result suggests that the effect of artificial bee colony algorithm is far more better than traditional searching algorithm. Artificial bee colony searching algorithm is able to find out the color image that satisfied to customer needs more accurately and effectively with a higher precision ratio and recall ratio.