Cartoon Image Content-based Retrieval System

With the rapid development of animation and new media industry, the growth of animation image data is beyond our image. How to rapidly and efficiently retrieval the desired images from large-scale animation image date becomes an urgent need, and an active topic in the research of image retrieval. The traditional image retrieval is unable to meet our need which is based on file name or the description of the image file. With the rapid growth of animation images, it is impossible to input all the animation image information manually. The result by retrieval animation images with descriptions is unsatisfactory. The purpose of retrieval animation image is to obtain the contents of the image information, so the result could provide the inspiration for the new animation creation. So animation images were retrieved by the content information of itself will greatly improve the efficiency and accuracy of animation image retrieval. The technology can promote the development of the animation industry.Although CBIR technology has been achieved some good results, such as fingerprint identification, face recognition, trademark retrieval, medical image processing, but not yet used in the animation image retrieval. In this dissertation, the basic principles and key techniques in the Content-based animation image retrieval technology are researched and analyzed, especially the feature extraction technology. Animation images are different from the landscape or figure images, the features of animation image was researched and analyzed in this dissertation. Color feature extraction which includes the color space quantization, color feature representation, color feature comparison of different images. Texture feature extraction which includes the LBP (Local Binary Pattern) texture feature extraction, the comparison of LBP texture features with different parameters, Tamura texture feature extraction.Finally, based on the key technologies of Content-based Image Retrieval, a development example of content-based animation image retrieval was presented. Detail needs analysis, design and retrieval processes of animation retrieval are depicted.