Multi-feature Image Retrieval Algorithm Based on Block Color Weighting
Zhang Ye · 2018
Aiming at the defects of traditional color histogram algorithm, this paper proposes a new weighted color histogram extraction algorithm. First, the image is processed as a block and according to the number of significant points in each block for the given weight, And then extract the color histogram features of each block separately, so as to obtain the color feature after image block weighting; While extracting images of Tamura texture, the color and texture of the two features together. Finally, the degree of similarity is reflected by the distance between the image feature vectors. The experimental results show that the proposed algorithm can effectively increase the spatial distribution characteristics when extracting eigenvectors, and achieve a better retrieve results.