基于主动形状模型和K近邻算法的人脸脸型分类

Face sharp classification based on active shape model and K nearest neighbor algorithm

针对人脸描述性脸型特征分类问题,提出一种新的基于主动形状模型和K近邻算法的脸型分类方法。根据主动形状模型方法定位得到的测试样本人脸边缘轮廓点,经归一化后以其围成区域面积作为人脸脸型特征。采用K近邻算法和面型指数实现测试图像的脸型分类。实验结果表明,该方法对人脸姿态变化有一定的鲁棒性,分类结果准确度高且脸型的分类符合人主观描述性判断。

Aiming at the classification problem of descriptive facial features, a novel face shape classification method based on active shape model and K nearest neighbor algorithm is proposed. Active shape model is used to extract facial eonlour points of the test sample, and then the enclosed area of the normalized facial contour points is regard as the facial shape features. Face shape classification of the test face image is realized by using K nearest neighbor algorithm and facial index. The experi mental results show that this method is robust for pose variation, the classification accuracy is high and classification results are consistent with subjective judgments.