Image Retrieval Technique Research Based On Texture And Shape Features

Content-based image retrieval is belonging to content-based visual information retrieval; it has got more and more attention in multimedia database management and date transmission. The image retrieval methods are mainly dependent on the features such as texture, color, shape, spatial relationship and so on. To begin with the texture and shape features extraction, this paper realizes the texture and shape feature-based image retrieval using co-occurrence texture model and directional detail histogram. The experiment results are given in the end. In this paper, the designed system for image retrieval using shape and texture features are first discussed.On the basis of researching the multimedia database and image retrieval technology,a new texture based and shape based image retrieval system between the multimedia database and the individual user is proposed. The retrieval system requires the user to provide some example pictures,the system will obtain the shape and texture features from the pictures automatically, and using these features to lookup the similar pictures from the database. The user can use the data from the result directly or begin a new query. Use the image retrieval technology based on texture and shape features can effectively improve the precision of image retrieval.The paper consist of five segments: The first segment mainly introduce the researching background of project,including the development of image retrieval technology and the development tools; The secong segment mainly design the structure of the retrieval system; The third segment mainly introduce the methods of obtain shape and texture features; The fouth segment mainly introduce the software design of the five modules in system; At the end of this paper, the appendix segment maily consist of source code of the functions referred in this paper.