Abstract: Aiming at the problem of face feature classification, this paper proposes a new face classification algorithm based on Active Shape Model(ASM) and K-nearest neighbor algorithm. It extracts feature points of face by ASM algorithm, normalizes all feature points, and computes Hausdorff distance between feature points and every sample of each class. The face is classified by K-nearest neighbor algorithm with the Hausdorff distance computed. Experimental results show that the algorithm has high classification accuracy and speed, and it is easy to realize.

Key words: face shape classification, Hausdorff distance, K-nearest neighbor algorithm, face feature extraction, Active Shape Model(ASM)

摘要： 针对人脸特征分类问题，提出一种基于主动形状模型(ASM)和K近邻算法的人脸脸型分类方法。将Hausdorff距离作为K近邻算法的距离函数，利用ASM算法提取待测图像的特征点，对点集进行归一化后计算人脸轮廓特征点与样本库中所有样本点集的Hausdorff距离，根据该距离值，通过K近邻算法实现待测图像的脸型分类。实验结果证明，该方法分类正确率高、速度快、易于实现。

关键词: 人脸脸型分类, Hausdorff距离, K近邻算法, 人脸特征提取, 主动形状模型

CLC Number: 

• TN911.73