Image Retrieval Model Based on Immune Algorithm

CBIR (content-based image retrieval) has become the main technique of image lib system. Its core is image similarity retrieval. The main obstacle of CBIR is that the retrieval effectiveness is unsatisfied. Since immune algorithm has ability of learning, memorizing and self-adapting in long term and in keeping with learning user's feedback information, it can improve the system recognition for users' semantic targets. Using excellence of immune algorithm, this paper proposes a new relevant feedback model based on immune algorithm and carries on the simulation tests for the above image retrieval model. The simulation indicated that the result of the beginning retrieving operation can meet the users' requirements very well and with more relevant feedback information the accuracy of the retrieving results are better.