摘要： 提出一种结合局部二值模式(LBP)和Fisherfaces的多模态人脸识别方法。用LBP算子提取人脸灰度图像和深度图像的区域LBP直方图序列(LBPHS)，再采用Fisherfaces分别构建相应的线性子空间，用余弦相似度作为投影向量的相似度量，用加权求和规则进行信息融合。在FRGC数据库上的实验结果表明，该方法要明显优于LBPHS与直方图交及Fisherfaces与余弦相似度的融合，等错误率仅为0.33%。

关键词: 局部二值模式, Fisherfaces方法, 多模态人脸识别

Abstract: This paper presents a method of multi-modal face recognition which combines Local Binary Pattern(LBP) with Fisherfaces. LBP descriptor is used to extract the LBP Histogram Sequence(LBPHS) from grey-level and depth face images. The corresponding linear subspaces are constructed by Fisherfaces respectively. The cosine similarity is adopted as the similarity metric of projected vectors. Weighted sum rule is utilized to fuse 2D and 3D information. The experimental results on FRGC database indicate that the Equal Error Rate(EER) of the proposed method is only 0.33%, which is much lower than fusion of LBPHS with histogram intersection and that of Fisherfaces with cosine similarity.

Key words: Local Binary Pattern(LBP), Fisherfaces method, multi-modal face recognition

中图分类号: 

• TP391.4