Abstract: Aiming at the characteristics of shoe sloes，a kind of slant correction algorithm is put forward.Based on this，a novel method by integrating distance histogram，angle histogram and Fourier descriptors is used to describe the shape feature that dynamic action leaves.According to the feature of shoe sloes，people’s characteristics，like the size of shoe and the pose of walking，can be mined.Finally，an experimental system is realized，and the result shows that the method proposed based on stable feature is fast，exact and practical.

Key words: recognition of shoe sloes, slant correction, stable feature

摘要： 针对鞋印图像自身特点，提出一种鞋印图像的倾斜校正算法；在此基础上提出用距离直方图、角度直方图与傅立叶描述子来共同表示动力作用留下的形状特征。根据提取的鞋印图像的特征可以进一步挖掘出人的某些特征，如脚的大小、走路姿势等。最后给出了一个实验系统，结果表明基于稳定特征的鞋印图像识别方法具有快速、准确、科学、实用等优点。

关键词: 鞋印识别, 倾斜校正, 稳定特征