Abstract: A novel image retrieval method based on region and relevance feedback is proposed in this paper.At first，the proposed approach employs a improved K-mean and fully unsupervised segmentation algorithm to divide images into regions，and low-level feature for the color，position，shape of each region are subsequently extracted and applied in similarity measurement.At last a relevance feedback mechanism，based on support vector machines is invoked to improve the retrieval performance.The experiment results show that the method can get better retrieval performance.

Key words: image retrieval, image segmentation, similarity measure, Relevance feedback, support vector machines

摘要： 提出了一种基于目标区域和相关反馈的图像检索方法，首先采用改进的K均值无监督分割方法将图像分割成区域，然后提取每个区域的颜色、位置、形状特征进行相似度计算；最后采用基于支持向量机（SVM）的相关反馈算法提高检索精度。实验结果表明，方法具有良好的检索性能。

关键词: 图像检索, 图像分割, 相似度计算, 相关反馈, 支持向量机