三维人脸识别算法研究

近年来,基于三维图像的人脸识别技术已经取得了很大进展,在约束环境下也能获得很好的识别性能,但仍受限于姿态、表情等因素,需要从算法上改进才能解决其影响。本文分别从基于空域直接匹配、基于局部特征匹配和基于整体特征匹配3个角度出发,对人脸匹配算法以及融合算法进行了研究,列出了部分改进算法的实验结果,并分析了算法有效性的原因,总结了目前面临的三维人脸识别算法难以突破的一些困难及未来的研究趋势。

Research on 3D Face Recognition Methods

The 3D image-based face recognition technology has made great progress in recent years,with good performance achieved under some constrained conditions.However,the technology is still limited by some factors such as facial pose and expression.To solve this kind of impact,the recognition methods must be improved.This paper shares the development status of 3D image face recognition by discussing a series of face matching methods and fuse-improve methods.And separately through three angles:spatial matching methods,local feature based methods and global feature based methods,to state some improvement of 3D face recognition methods.Besides,some experimental results of improved methods are listed and the reasons about effectiveness of the methods are analyzed.Finally,the paper summarizes some challenges which hinder the improvement of 3D face recognition methods,and explore the future research trend.