基于HOG和特征描述子的人脸检测与跟踪  

Face detection and tracking based on HOG and feature descriptor

作　　者：李澎林[1] 邹嘉程 李伟[1] LI Penglin;ZOU Jiacheng;LI Wei(College of Computer Science and Technology,Zhejiang University of Technology,Hangzhou 310023,China)

机构地区：[1]浙江工业大学计算机科学与技术学院,浙江杭州310023

出　　处：《浙江工业大学学报》2020年第2期133-140,共8页Journal of Zhejiang University of Technology

摘　　要：在人脸检测和跟踪过程中,真实场景的光照变化和随机噪声会降低检测准确率,而多张人脸的存在和人脸的姿态变化会影响单个目标的跟踪。针对这些问题,提出了一种基于HOG和特征描述子的人脸检测与跟踪方法。在人脸检测过程中,利用方向梯度直方图(HOG)特征来检测视频帧中的人脸,提高了检测的准确度;在人脸跟踪过程中,采用了一种结合特征描述子的跟踪校正策略,利用基于欧氏距离的方法进行人脸相似度对比,并以此更新跟踪结果,降低了多人脸因素的干扰。实验结果表明:笔者算法的人脸检测与跟踪准确率较高,鲁棒性较好。In face detection and tracking process,illumination changes and random noise of real scene will reduce the accuracy of detection.Moreover,presence of multiple faces and face posture changes will affect the tracking of a single target.To deal with these problems,a kind of face detection and tracking method based on HOG and feature descriptors is proposed.In the face detection process,the histogram of oriented gradient(HOG)feature is used to detect faces in video frames,which improves the accuracy of detection.In the face tracking process,a tracking correction strategy combined with feature descriptors is adopted.The strategy based on Euclidean distance is used to compare similarity of faces,and the comparison result is used to update tracking results.This will reduce the interference of multi-face factors.The experimental results show that the proposed algorithm has high accuracy of face detection and tracking and good robustness.

关 键 词：人脸检测 人脸跟踪 HOG 特征描述子 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]