基于人脸分块近似对称性预处理的人脸识别算法  

Face Recognition Algorithm Based on Face Segmentation Approximate Symmetry Preprocessing

作　　者：朱叶 李琳[1] 张佳鹏 ZHU Ye;LI Lin;ZHANG Jia-peng(College of Optoelectronic Information and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200082,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200082

出　　处：《软件导刊》2021年第4期75-80,共6页Software Guide

摘　　要：针对非限制场景下局部遮挡带来的准确率不高、稳定性差等问题,提出一种基于人脸分块近似对称性预处理的人脸识别算法。根据人脸特征点定位,将人脸分成左眼区块、右眼区块、左鼻区块、右鼻区块、左嘴区块、右嘴区块,将剪裁好的6个人脸区块与背景区块进行遮挡判断,若判断出来的人脸区块为遮挡区块,则根据人脸近似轴对称,采用迭代梯度下降算法恢复出完整的人脸图像。恢复后的人脸图像用SRC算法进行人脸识别。实验结果表明,在有遮挡的人脸识别情况下,该算法识别率可达97.6%,基于人脸分块近似对称性预处理的人脸识别算法比其它算法鲁棒性好,识别率有所提高。To solve the problems of low accuracy and poor stability caused by local occlusion in unrestricted scenes,a face recognition algorithm based on approximate symmetry preprocessing of face blocks is proposed.According to the location of facial feature points,the face is divided into six blocks:left eye block,right eye block,left nose block,right nose block,left mouth block and right mouth block.The cut six face blocks and background blocks are used to judge the occlusion,if it is judged that the face block from cues is an occlusion block,then iterative gradient descent algorithm is adopted according to the approximate axial symmetry of the face to recover to produce a complete face image.The recovered face image is recognized by SRC algorithm.Experimental results show that the recogni⁃tion rate of face recognition with occlusion can reach 97.6%.Compared with other algorithms,face recognition algorithm based on ap⁃proximate symmetry preprocessing of face block has better robustness and higher recognition rate.

关 键 词：人脸识别 非限制场景 局部遮挡 人脸分块 近似对称性 特征点定位 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]