:人耳识别系统中,需要根据特征提取的要求实现图像的自动归一化,并且选择合适的依据使归一化后的结果具有一定的有效性和合理性。提出了一种新的外耳长轴标记法,根据外耳长轴将人耳统一到标准的尺寸和角度 运用主动形状模型算法,实现了外耳轮廓的自动搜索,且对算法进行了改进,调整了模型的初始位置。实验证明,改进后的主动形状模型(ASM)能更快速准确地收敛,且归一化的方法能够合理地将人耳标准化,为后续的特征识别提供了保障。-In the ear recognition system, it is required to realize the automatic normalization of the image according to the requirement of the feature extraction, and to　the appropriate basis to make the results of the normalization has a certain validity and rationality. A new auricle major axis labeling method is proposed. According to the long axis of the outer ear will ear into a unified standard size and angle　using active shape model algorithm, the realization of the automatic search of the outer ear contour and the algorithm was improved to adjust the initial position of the model. The experimental results show that the improved active shape model (ASM) can converge more quickly and accurately, and the normalization method is able to provide the guarantee for the subsequent feature recognition.