Research On Color Feature Based Image Retrieval

Content based image retrieval is one of the hottest spots now.Color is the easiest and most important character. By reason of the disadvantages of block-color histogram,the paper find a new method to improve it.In this thesis, our research work mainly focuses on the following aspects:1. Gives the different quantification progression to the different color channel,then composing one dimension vector with sixty three dimension vectors,in order that we can make the algorirhm have higher efficiency .2. Takes the overalapped method to get the space information.3. Form a mathematics model of match question.To solve the problem,we take a classic method——Hungary algorirhm to calculate the likelihood between pictures.All these make the algorirhm stronger.At last I make a system to prove the algorirhm.