Content-based Image Database Retrieval System

With the rapid development of Internet and multimedia technology, the sources and the numbers of images continue to expand. The manners of Text-based retrieval and Folder-based storage management have been unable to meet the people's demands for images processing. How to store and retrieve these image resources much more effectively had put forward a new challenge to the majority of researchers. Therefore, content-based image database retrieval systems gradually become the research focus.In the thesis,an image database system of content-based retrieval is designed and established. This system is composed of two modules which are image database module and image retrieval module.The image database module is used to store not only the original images, but also the image features after extraction which is established by Microsoft Access. And the original images can be automatically classified according to some types of themselves.The image retrieval module not only can be used to perform the operations such as image storage, display, delete, and add description, but also can be used to perform keyword-based image retrieval and content-based image retrieval. Firstly, the color, texture and shape features are extracted by gray-scale histograms, gray co-occurrence matrixes and the boundary feature extraction methods and stored in Access image database. Secondly, text-based search and thumbnail browsing are used to locate the example images from the Access database. The example images are used to perform the content retrieval and matching based on color features, texture features, shape features and integrated features. At last the analysis and evaluation are given to estimate the retrieval results.The experiment results indicated that the retrieval system design idea is effective and feasible. The research results can be given more comprehensively and accurately.