Content-based Image Retrieval Technology Research And Application

With the wider use of informational & digital technology and multi-media technology and the rapid developments of computer web, a great amount of information from all kinds of industries are digitted in the method of multi-media technology. Image is the multi-media information, which is the most basic and used whidely. It has become a pattern of popular digital information. The application of image data has been linked to every aspect of scientific technology and daily lives. Under the background of this kind, a lot of various image data have generated, which requires us to manage them scientifically. Therefore people do some research for the field of image recalling. Today image recalling has become the hit of research for computer.Content-based image retrieval (CBIR) opens a way for recalling the image from visual direction more accurately and visually. At present, content-based image retrieval uses the visual contents such as color, texture, shape, spatial relationship to indicate and recall image. Content-based image retrieval is a very active studying problem; many special topics on this almost appeared many international famous periodicals. However, generally speaking, the research for content-based image retrieval technology is not incomplete, which is difficult to reach the practical standard. Therefore, there's not only important theoretical sense but also practical value to strengthen this kind of research. There are three keys to image-recalling basing on contents. First, choosing the right image features. Second, taking some useful measures to get obtain the features. Third, having the accurate algorithm, which matches the features. The aim of this passage is to select the suitable features, improve the measures to obtain the features and use the reasonable algorithm which matches the features and recall the image in order to build up brilliant & exact system of image-recalling.This paper illustrates the current generalizations of developments of the image-recalling basing on the contents, which analyzes the features of the recalling system basing on contents, the important role of it and the problems of it and does some research for the structure features and key technologies of CBIR system. The paper makes further research for the technologies in processing image features, using the color, texture and shape to carry out the recalling basing on contents, improving some math's problems, analyzing the experimental results, drawing the useful conclusion; giving the experimental system of image-recalling and building up the image data memory. The users enter the images in advance, when searching can obtain the images and demonstrate. In order to improve the speed of search the features memory is established and is related to image memory, so only needing calculating the shown images' feature value when carrying out the similar comparison. The feature value of the images in the memory can be drawn directly for the feature memory, which saves a lot of time.