多视角三维人脸识别中的特征提取

针对二维人脸识别系统在姿态发生较大变化就难以识别的问题,利用三维人脸数据姿态不变性的特征,提出了一个基于方向最大值的方法来估计鼻尖点,同时也给出姿态的角度.用子空间表示的鼻子轮廓模型被用来选择鼻尖点的最可行的候选点.另外,利用SUSAN算子提取边缘,并与方向积分投影等方法结合,快速准确的定位内外眼角点和嘴角点.实验证明该方法可以保证达到和手工自动标定特征点相差无几的准确率.

Feature Extraction in Multiview 3D Face Recognition

Aiming at questions that 2D face recognition systems have encounter difficulties in recognizing faces with large pose varia- tions, utilizing the pose- invariant features of 3D face data ,a method based on the directional maximum is proposed to estimate the nose tip location and the pose angle simultaneously. A nose profile model represented by subspaces is used to select the best candi- dates for the nose tip. Buy using SUNSAN operator to extract the corner points and combined with directional integral projection. Ex- perimental results show that this method provides recognition accuracy that is comparable to the accuracy of a system with manually labeled feature points.