The Research Of Image Retrieval Based On Edge And Color Feature

With the rapid development of computer and multimedia technology, more and more multimedia information come forth. The research of image retrieval algorithm based on single features like color, texture and edge has achieved good performance. But, the result of image retrieval algorithm only relies on single features is not very good. In the image retrieval algorithm based on content, edge is the most basal feature of image. It contains abundant information. The color feature has the advantage of calculating easily and good robustness. As an important vision information has attend extensive application.This paper researches on the feature extraction algorithm based on color, analyses the shortage of existing algorithm and proposes an advanced algorithm. The advanced algorithm extracts the color feature of edge area for image retrieval. The perfect aspect is dividing the edge into inner edge and outer edge. Increase the weight of the inner edge (target part). This is realized through the method of the multi-scale and multi-structuring elements mathematical morphology. The algorithm by setting a weight takes the important of the target into account and overcomes the shortcoming of color histogram which loses the location information.This paper analyses the algorithm based on edge and color features. When extracting the edge and color features, those algorithms use the same parting style. This ignores the particularity of edge and color greatly. In the advanced algorithm using different parting style in edge and color features, Consider both edge and color particularity and attend good retrieval results.Finally, construct a content based image retrieval system. Do experiment on the algorithm proposed in this paper. The experimental results show that those algorithm effectively improve the image retrieval performance...