An Optimized Image Retrieval Method Based on Hierarchal Clustering and Genetic Algorithm
Min Huang, Bo Sun, Jianqing Xi · 2009
Image search on Web is very familiar to various users, and improving the efficiency and accuracy of image search has become more and more a hotpot in this research field. For different commercial image engines use different retrieval techniques respectively, the coverage area and accuracy of each individual search engine await development. An improved method based on multi-optimization techniques of image retrieval is presented in the paper. On the base of relevance feed-back principle, the method does some work of the vectorization and weights adjusting to the images generated by commercial image engines, and then adopts hierarchal clustering and genetic algorithm techniques to optimize the results further. Finally, by developing a prototype of image retrieval engine based on the method presented and doing some tests, the advancing in accuracy of image retrievals of the method has been proved.