Abstract: To express the information for image more accurately and improve the performance of image retrieval, an improved algorithm is proposed based on Integrated Region Matching（IRM）. The image is segmented using threshold and fuzzy C-means algorithm. Then IRM algorithm is used to match the image which is improved on integrated region distance and important factor, and a concept of the effective distance is proposed according to the percentage relationship of objective and background. The experimental results show that the average precision of the proposed algorithm is raised by 4.58% compared with the classic algorithm. The method can be widely used in the image retrieval system, which has greater practical value.

Key words: image segmentation, integrated region matching, effective distance, image retrieval

摘要： 为了能更准确地表达图像信息，提高系统检索性能，提出了一种基于综合区域匹配（IRM）的改进算法。先采用阈值和模糊C-均值相结合的方法分割图像，再采用改进了综合区域距离和重要性因子算法的IRM方法进行图像匹配，并根据图像目标和背景的面积比关系提出“有效距离”概念。实验结果表明，相对于经典算法，改进后算法的平均查准率增加了4.58%。该方法能广泛应用于图像检索系统，具有较大适用性。

关键词: 图像分割, 综合区域匹配, 有效距离, 图像检索