基于局部径向二值模式的三维人脸识别

针对人脸识别中局部特征的提取,提出了局部径向二值模式(LRBP,Local Radial Binary Pattern),并将其用于三维人脸识别.首先,对经过预处理的人脸深度图像进行区域划分;然后用局部径向二值模式提取子区域的特征序列,并将其链接在一起构成三维人脸的特征向量;最后,利用Fisherface方法对三维人脸特征向量进行训练和识别.在中国科学院自动化研究所三维人脸数据库中选取样本,利用LRBP对其进行识别,结果表明该方法在基本不损失识别率的前提下,可以有效提高识别的效率.

3D face recognition based on local radial binary pattern

An operator named local radial binary pattern (LRBP) was proposed for extracting local features in face recognition. The binary sequence encoding scheme of the LRBP is different from that of the local binary pattern. Firstly, the proposed LRBP operator was used in 3D face recognition. 3D face depth images were preprocessed and divided into subregions. Then the signature sequences of the subregions were extracted by the LRBP operator. The feature vectors of a 3D face depth image were obtained by connecting the signature sequences of all the subregions of the image. Finally, the 3D face feature vectors were trained and recognized using the Fisherface method. Experiments were conducted using the 3D face database of Institute of Automation, Chinese Academy of Sciences. The results show that the proposed method can effectively promote the efficiency of 3D face recognition without reducing the recognition rates.