基于脊点特征的三维遮挡人脸识别方法  

Recognition Method for 3D Occluded Face Based on Ridge Point Features

作　　者：王丽娟[1] 王洁浩 魏一帆 WANG Lijuan;WANG Jiehao;WEI Yifan(School of Electronic Information Engineering,Xi'an Technological University,Xi'an 710021;Xi'an Branch,Shenzhen Huichuan Technology Co.,Ltd.,Xi'an 710003)

机构地区：[1]西安工业大学电子信息工程学院,西安710021 [2]深圳汇川技术股份有限公司西安分公司,西安710003

出　　处：《计算机与数字工程》2025年第1期103-107,163,共6页Computer & Digital Engineering

摘　　要：针对当前三维人脸遮挡识别困难的问题,提出一种有效的基于脊点特征的三维遮挡人脸识别方法。主要基于三维遮挡人脸的脊点模型,先利用人脸各个被遮挡区域脊点的空间分布特点来检测遮挡区域,再采用3D形状上下文描述子(3D Shape Context)表征未遮挡区域脊点点集,和库中人脸模板的相应区域脊点点集进行匹配,并为各区域的匹配结果赋予不同权重,求和得到整体匹配结果。在Bosphorus和UMB-DB数据库上对论文方法进行验证,实验结果表明该方法可以有效检测出遮挡区域,并能够有效识别三维遮挡人脸,且识别准确率高。A simple yet effective method for the recognition of occluded 3D face is proposed in this paper.The method is main⁃ly based on the ridge points model of the 3D occluded face.The spatial distribution of the ridge points of each occluded region of the face is used to determine whether the region is occluded or not.After that,the ridge point sets of the un-occluded region are repre⁃sented by 3D Shape Context descriptor and compared with the ridge point sets of the corresponding region of 3D face template.Differ⁃ent weights are assigned to the matching results of each region,and the overall matching results are obtained.The method is tested on the Bosphorus and UMB-DB database.The results show that the occluded area can be detected effectively and the 3D occluded face can be effectively recognized with high accuracy.

关 键 词：三维人脸识别 遮挡 脊点 形状上下文 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]