三维人脸识别研究进展综述  

Overview of research progress on 3-D face recognition

作　　者：罗常伟 於俊[2] 于灵云 李亚利[1] 王生进[1] LUO Changwei;YU Jun;YU Lingyun;LI Yali;WANG Shengjin(Department of Electronic Engineering,Tsinghua University,Beijing 100084,China;Department of Automation,University of Science and Technology of China.Hefei 230027,China)

机构地区：[1]清华大学电子工程系,北京100084 [2]中国科学技术大学自动化系,合肥230027

出　　处：《清华大学学报（自然科学版）》2021年第1期77-88,共12页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金面上项目(617711288)。

摘　　要：近年来,三维人脸识别研究取得了较大进展。相比二维人脸识别,三维人脸识别更具有优势,主要特点是在识别中利用了三维形状数据。该文首先根据三维形状数据的来源,将三维人脸识别分为基于彩色图像的三维人脸识别、基于高质量三维扫描数据的三维人脸识别、基于低质量RGB-D图像的三维人脸识别,分别阐述了各自具有代表性的方法及其优缺点;其次分析了深度学习在三维人脸识别中的应用方式;然后分析了三维人脸数据与二维图像在双模态人脸识别中的融合方法,并介绍了常用的三维人脸数据库;最后讨论了三维人脸识别面临的主要困难及发展趋势。Research on 3-D face recognition has made great progress in recent years.3-D face recognition is more effective than 2-D face recognition.Its main feature is the use of 3-D shape data for recognition.The 3-D face recognition methods are categorized into three types based on the source of the 3-D shape data with methods based on 2-D color images,high quality 3-D scanning data,and low quality RGB-D images.This study reviews these methods and discusses their advantages and disadvantages.This paper then reviews the use of deep learning methods for 3-D face recognition.Besides,3-D and 2-D face data fusion methods are reviewed for bi-modal face recognition.The commonly used 3-D face databases are also summarized.Finally,the main difficulties and future development trends of 3-D face recognition are discussed.

关 键 词：三维人脸识别 三维数据 深度图像 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]