Index-based fast search algorithm of image database on internet

For the application over internet, a multimedia system should efficiently manage various types of data such as text, image, audio, and video and quickly increasing number of users makes network bandwidth more and more precious. In this paper, we propose a novel algorithm for fast image database searching over internet. In the proposed system, the search results first returned are returned the rough images and the user can preview and narrow down to the images he wants. By Interactive feedback, the user can obtain the more accurate results by refining. We present some experiments to measure it's the efficiency of the proposed algorithm. In addition, the quality of query result is about 1 dB better than the JPEG coded image at the same bit rate. Based on these results, our search algorithm is an effective retrieval method for image database.