解决人脸识别光照问题的预处理方法综述

A Survey Preprocessing Methods for Illumination Insensitive Face Recognition

光照问题是人脸识别面临的一个难题,它在很大程度上影响着人脸识别系统的性能。近年来人们提出了很多解决方法,其中预处理方法因其简单高效与实用而备受关注。本文结合近年来已发表的文献,按照经典图像处理方法(直方图变换和伽马亮度校正GIC)、结合朗伯表面的方法(自商图和局部规格化)和基于图像分解的方法(离散余弦变换和经验模态分解),对预处理方法在解决人脸识别光照问题中的应用进行了回顾、比较和总结。

Illumination variation is one of the challenging problems in face recognition, which can greatly affects the performance of face recognition systems. In the past few years, various methods have been proposed to handle the illumination problem. The preprocessing approaches commonly exploit simple and efficient image processing techniques so that they have been given more and more attention. The typical preprocessing approaches can be categorized into classical methods （histogram processing and Gamma intensity correction）, Lambertian surface-based methods （self-quotient image and local normalization） and image decomposition-based methods （discrete cosine transform and empirical mode decomposition）. This paper presents some review and summaries about those typical preprocessing approaches.