Content Based Image Retrieval Based On Color And Texture Feature

With the rapid development of Internet and multimedia technologies, a large amount of digital image information store on the Internet. Image is an abundant amount of information bearing carrier, which is widely used in various industries. In order to extract the desired image from a large number of images quickly and accurately,Content-based Image Retrieval appeared in the 1990 s. The idea is that the image contains color, image texture features, shape and spatial relationship to describe the image, these features as an index, retrieve similar or related images.The paper aims to search color and texture features of image retrieval mainly,integrated image retrieval method is proposed a fusion of color and texture features.Considering the amount of calculation and conversion between RGB color model, as well as with human visual perception, the HSV color model is closer to real color image information in the color feature extracting. In order to solve the problem that color histogram can not express the spatial distribution, algorithm is proposed, which is based on making blocks and primary color matching algorithm image retrieval. The image is divided into blocks, each block to extract the dominant color as the color feature. In the texture feature extracting, the Gray Level Co-occurrence Matrix texture is widely used in describing the image. And texture feature extracting adds to the division of information on the frequency domain filter for image enhancement on this basis, and then converted to take advantage of the spatial domain Gray level co images calculated value of the four characteristic image as a texture feature. Color and texture features weight fusion as a final retrieval features.1. In-depth analysis of the image content as the core search technology, including an image(color,shape,texture features) extraction algorithm, and the similarity between the image matching criteria, choose the most suitable color model,after the analysis of color model. system is more availability and efficiency.2. Single feature can not achieve satisfactory results. Thought multiple feature fusion has begun to apply to a specific search, the paper will combine color and texture features,which is weighted feature fusion for image retrieval. The experimental results,the proposed algorithm this paper shorten the retrieval time, comparing the results of the algorithm.