Adaptive Feature Integration Image Retrieval Research

With the popularization and development of Internet, the number of image data grows dramatically fast, and that how to retrieve image efficiently and quickly becomes an important issue in the field of image's application. In order to manage and retrieve large amounts of images, the CBIR has emerged to be one of the hot research areas in image process domain.Based on the analysis of the CBIR systems in existence, for the problems that the actual CBIR systems can not distill effectual and accurate feature combination , the dissertation do some researches as follows:Firstly the history and development of CBIR are introduced, and the algorithms of CBIR which based color, texture, and shape feature are discussed respectively. For redeeming the missing space information, a new method of partition color histogram is presented, and the paper compared the algorithm with the traditional color histogram by experimentation. Experimental results have shown that the partition color histogram method has better performance.The method which based on single feature can only express part of image information, only describe exparte content of the image and lack of distinct information, a method image retrieval method based color, texture and shape feature is researched and the method is compared with the based on single feature method by doing research. After studying the results, the dissertation summarized merits and shortcomings of the method and suggestions for improvement are put forward. By discussing the shortcomings of the static feature selected image retrieve method, a adaptive feature selection image retrieval method is presented. Based on the analysis of the image features, this method can adjust the weights of the feature in feature combination, in order to improve the effect of the CBIR without user's helping. A large number of experiments show that the adaptive feature integration method is better than the tradition static feature selected image retrieve method.Finally, the work of this paper is summarized, and the forecast of the development of content-based image retrieval is discussed.