本文基于双树复数小波变换的多尺度、多方向选择特性，对Jim Mutch 提出的视觉信息处理模型进行改进。改进后算法大大提高了运算速度，同时获得的特征向量具有尺度、旋转和平移不变性。最后，利用支持向量基进行分类识别。Caltech101 人脸数据库的实验结果表明，该方法能够取得较好的分类效果。关键字：视觉信息处理；HMAX 模型；支持向量机Abstract: This paper advances Jim Mutch’s proposed model, based on the multi-size and multi-orientation selecting trait of dual-tree complex wavelet translation. It improves the speed. And feature vector, that is size-invariant, rotating invariant and position invariant, can be obtained. Finally, support vector machine is used to classify. The result of the experiment on Caltech101 face images indicates that the proposed method has good classification performance.Key words: visual information processing; HMAX model; support vector machine