Content-based image retrieval is a research hotspot recently. Image mining applies classification of data mining, which improve the efficiency of CBIR. The terminal user of CBIR system is human being, thus it is important to capture image knowledge from the point view of psychics. Relevance feedback is used for CBIR in order to embed user model into image search.K-Nearest Neighbor Algorithm is a typical classification algorithm. In this paper author applies an algorithm: relevance feedback k-nearest neighbor algorithm, utilizing image by key word image search. The algorithm improved the classification efficiency and the insufficient of accuracy which is related with the number of samples. Moreover, author implements a simulation system to test the performance of author's algorithm. The result of experiment with relevance feedback shows that author's algorithm can improve precision of image mining. |