基于样例学习的面部特征自动标定算法

Example-Based Learning for Automatic Face Alignment

面部特征标定是人脸识别中的一个关键问题.提出了一种基于样例学习的面部特征自动标定(人脸形状自动提取)方法.该方法是基于下面假设提出来的:人脸图像差和形状差之间存在一种近似的线性关系——相似的人脸图像在较大程度上蕴涵着相似的形状.因此,给定标注了特征点的人脸图像学习集,则任意新的输入人脸图像的面部形状可以采用如下方法估计:测量该人脸图像和训练集中图像的相似度,并将同样的相似度用于该人脸图像形状的重建.即:如果输入人脸图像可以表示为训练图像的优化的线性组合,那么同样的线性组合系数就可以直接用于训练集对应形状的线性组合从而得到输入人脸图像的形状.实验表明,该算法相对于其他传统的特征标定算法具有可比的精度和较快的速度.并且,还将此算法扩展到了多姿态情况下,实现了多姿态人脸图像形状的自动提取.

A novel example-based automatic face alignment strategy has been proposed for facial features alignment, i.e. facial shape extracting. The method is motivated by an intuitive and experimental observation that there exists an approximate linearity relationship between the image intensity difference and the shape difference, that is, similar face image intensity distribution implies similar face shape. Therefore, given a learning set of face images with their corresponding face landmarks labeled, the shape of any other face image can be learned by estimating its similarities to the training images in the learning set and applying these similarities to the shape reconstruction of the unknown face image. Concretely, if the unknown face image is expressed by an optimal linear combination of the training images, the same linear combination coefficients can be directly applied to the linear combination of the corresponding training shapes to construct the optimal shape for the novel face image. Experiments have shown that, compared with traditional methods, the proposed method can achieve a comparable alignment accuracy in less time. Furthermore, the same strategy has been extended to extract the shape of face images with varying poses.