Image retrieval based on blocked histogram and Sobel edge detection algorithm

This paper proposes a new of image retrieval based on blocked histogram and Sobel edge detection and the metergrowth matrix of gray scale. This algorithm in color feature draw in first the quantification in color and the transformation for color space after, carry out a kind of new overlapping minute again for image piece, this minute has outstanding piece, then, for the retrieval of goal area, draw the color histogram that divides into piece as the color feature of image, show via experiment, should divide into piece in for image retrieve precision higher. In veins feature draw in for the retrieval speed that raises image, first with Sobel operator, draw image edge, then apply the metergrowth matrix of gray scale to draw the veins feature of edge image. Experiment result proves that what this paper put forward make up color with the image retrieval of veins feature is better than unitary image retrieval have better check whole rate with check accurate rate.