Abstract: With the rapid development of Internet technology，image documents have become an important information source.It is hard to retrieve certain images from all available ones.This paper proposes an interactive image recommendation system，which uses color histogram feature or Gabor texture feature to express image contents，then a kernel based K-meanse is utilized to cluster images into multiple classes by their visual features，finally based on a sample and hyperbolic techniques，images are recommended and displayed.The experimental results demonstrate that the proposed system can recommend and display the similar images from the same class efficiently，when users click on the images they are interested in.

Key words: image recommendation, similarity-preserving image projection, hyperbolic

摘要： 随着互联网信息的快速发展，图像信息已经成了一个重要的信息来源，但是对用户来说很难找到自己需要的特定的图像，构建了一个交互式的图像推荐系统，用颜色直方图特征或者Gabor纹理特征来表示图像的内容，根据图像的视觉特征采用一种基于核的K均值聚类方法将图像聚成多类，采用hyperbolic显示技术，将图像按照视觉特征显示在屏幕的不同区域。结果显示该方法可以有效快速地帮助用户在图像数据库中找到自己需要的图像。

关键词: 图像推荐, 相似度保持投影, hyperbolic显示