Multimedia Content-based Image Retrieval

In traditional method of image retrieval searches images according to keys, number and sort describe from image database.Along with the development of multimedia technology and Internet,the image resource of Internet increasingly add, the traditional text-based retrieval can not meet the require of people to retrieval and manage these image resource. To account for the deficiency of text-based retrieval, content-based visual information retrieval was been bring forward and get attention under the background of visual information at very fast speed expand.Content-based image retrieval, belong to the one research area of the image analysis. Its English glossary called CBIR, Content-based image retrieval, or content-based visual information retrieval. The purpose of Content-based image retrieval is to search image in the image's database and find out image accord with query term under precondition of require. Its include two steps:●Feature Extraction: The first step in the process is extracting image features to a distinguishable extent.●Matching: The second step involves matching these features to yield a result that is visually similar.So that, Content-based image retrieval according as hominine's vision characteristicisticistic, distill the original vision information from image, mostly include the color, texture and shape feature of the image. Then use characteristicistic matching arithmetic to searching and matching. CBIR system r is mostly according as image vision character to scale the comparability between images. Characteristicistic was choose to stand or fall, straight influence of search's succeeded. The capability and speed of search system was decided by characteristicistic distill arithmetic. So, the system kernel of Content-based image retrieval is choose and distill the characteristicistic of image.This text puts forward a kind of method of the color coherence vector based retrieval, increasing biggest inspectional accuracy. And reduce the light feels strength to analytic interference in characteristic in image. Pass to experiment, adopt this method, can obtain the good experiment result. In Vision-based characteristic distill technique, to analysis in common use of image segmentation method, this text adoption Median-wave-based and mathematics-appearance image edge examine technique. This technique can nicely wipe off the voice of yawp, examination details in the image, position accurate, the consecution is better, and apt to the program software.