本文已被:浏览 2700次 下载 4704次
Received:May 12, 2010 Revised:June 06, 2010
Received:May 12, 2010 Revised:June 06, 2010
中文摘要: 目前,光照问题已成为影响人脸识别率的关键因素,对人脸识别中处理光照的常用方法进行了对比实验,实验结果表明,光照锥法及近似光照锥的各种算法对光照的明显变化都具有较强鲁棒性,而其他算法在光照变化明显情况下,识别率都有明显的降低,对处理三维人脸图像光照的光照锥和球谐理论的算法,进行了全面深入研究,分析了它们的原理、优势、劣势和结果,提出了以后的研究工作的方向。
Abstract:At present, Light has become a key factor of affecting face recognition rate, the common method of deal wih illumination were compared in face recognition, experimental results show that the illumination cone method and the approximate light cone light methods have strong robustness when have significant illumination changes. while the other algorithms in the changing illumination cases, the recognition rate has significantly decreased. Deal with three-dimensional face image algorithm, illumination cone algorithm and the spherical harmonic theory have been elaborated. Extensive in-depth efforts have been put into the research toward illumination treatment methods in face recognition. Their strategies, advantages/disadvantages and performances are elaborated. Several promising directions for future research have been suggested.
keywords: face recognition wavelet decomposition gabor filter illumination cone spherical harmonic theory
文章编号: 中图分类号: 文献标志码:
基金项目:基金项目:成都理工大学研究基金(2008YQ01)
引用文本:
刘笃晋,孙淑霞,李思明.人脸识别中光照处理方法的分析.计算机系统应用,2011,20(1):160-
LIU Du-Jin,SUN Shu-Xia,LI Si-Ming.Analysis of Illumination Treatment Methods in Face Recognition.COMPUTER SYSTEMS APPLICATIONS,2011,20(1):160-
刘笃晋,孙淑霞,李思明.人脸识别中光照处理方法的分析.计算机系统应用,2011,20(1):160-
LIU Du-Jin,SUN Shu-Xia,LI Si-Ming.Analysis of Illumination Treatment Methods in Face Recognition.COMPUTER SYSTEMS APPLICATIONS,2011,20(1):160-