基于肤色与特征的视频流人脸快速检测与跟踪方法研究及系统实现
|
摘要
信息数字化和网络化的高速发展对国家以及社会生活安全提出了全新的要求。在这种环境下,传统的安全技术显得力不从心。而生物特征认证技术是解决信息化、数字化、网络化社会安全问题的最好的办法之一。常用的生物特征认证技术包括:人脸识别、虹膜识别、指纹识别、声音识别、步态识别等。在众多的生物特征认证技术中,人脸识别因具有获取直接隐蔽、识别速度快、识别准确率高、安全性高、使用条件简单,非侵犯性等特点,在公安、安防、银行、计算机网络信息安全等诸多领域内具有广阔的应用。
人脸识别研究工作一般分为三个过程:人脸检测,面部特征定位以及人脸特征提取与识别。而人脸检测是人脸识别过程中最为基础和关键的环节。只有精准的定位人脸的位置,才能够准确的识别人脸。
本文首先概述了目前国内外比较流行的几种人脸检测的算法,分析了各种算法的优缺点,结合人脸的肤色和特征几何分布特征,我们提出了基于肤色与特征相结合的视频流人脸快速检测的方法,并建立了系统原型,适用于实际网络在线考试中的身份认证。本文以普通PC摄像头作为图像采集设备,以采集的视频流为数据源,截取视频流中的单帧图像,通过转换彩色空间、人脸肤色建模、后处理操作实现了人脸定位,通过图像马赛克、边缘提取,显著特征点定位等技术定位人脸图像的各特征点。在此基础上通过人眼定位实现了在视频流中对于人脸的跟踪的算法研究。
试验结果表明,我们所实现的人脸检测适用于近距离人脸的检测,定位速度快,误检率低,可以应用于网络考试的身份认证。
The rapid evolutions of digital information and Internet rise new demands for safety of country and social life. In this situation, traditional safety technologies show incompetent. Biological authentication technology is one of the best ways to solve the safety issue appearing with the development of digital information and Internet. Among these technologies, because of the merits such as directness, quick recognition, high performance, high safety, easy use, and no impingement, face recognition has many applications in the fields of police, safety protection, bank, and net information safety.
The process of face recognition includes three steps: face detection, locating characteristics of face, face characteristics extraction and recognition. Among them, these detection is the most fundamental and important one.
This paper researches on some popular algorithms inside and outside of our country. Meanwhile, these algorithms are comparably analyzed about their merits and shortcomings. We rise a rapid detection and tracking algorithm based on the color and character of human face. A prototype of the system applicable to the examination of the actual online network authentication. In this paper as an ordinary PC camera image acquisition equipment to collect the data source for streaming video, video streaming interception in the single-frame images, through the conversion of color space, color modeling Face, after handling Face positioning realized through Image mosaics, edge extraction, notable feature points, and the setting of technical positioning of the Face Image feature points. On the basis of this position through the eyes achieved in video streaming in the face tracking algorithm research.
Test results show that we are realizing the face detection and tracking algorithm applied to the face at close range detection, positioning speed, low error rate seized, can be used to test network authentication.
人脸识别研究工作一般分为三个过程:人脸检测,面部特征定位以及人脸特征提取与识别。而人脸检测是人脸识别过程中最为基础和关键的环节。只有精准的定位人脸的位置,才能够准确的识别人脸。
本文首先概述了目前国内外比较流行的几种人脸检测的算法,分析了各种算法的优缺点,结合人脸的肤色和特征几何分布特征,我们提出了基于肤色与特征相结合的视频流人脸快速检测的方法,并建立了系统原型,适用于实际网络在线考试中的身份认证。本文以普通PC摄像头作为图像采集设备,以采集的视频流为数据源,截取视频流中的单帧图像,通过转换彩色空间、人脸肤色建模、后处理操作实现了人脸定位,通过图像马赛克、边缘提取,显著特征点定位等技术定位人脸图像的各特征点。在此基础上通过人眼定位实现了在视频流中对于人脸的跟踪的算法研究。
试验结果表明,我们所实现的人脸检测适用于近距离人脸的检测,定位速度快,误检率低,可以应用于网络考试的身份认证。
The rapid evolutions of digital information and Internet rise new demands for safety of country and social life. In this situation, traditional safety technologies show incompetent. Biological authentication technology is one of the best ways to solve the safety issue appearing with the development of digital information and Internet. Among these technologies, because of the merits such as directness, quick recognition, high performance, high safety, easy use, and no impingement, face recognition has many applications in the fields of police, safety protection, bank, and net information safety.
The process of face recognition includes three steps: face detection, locating characteristics of face, face characteristics extraction and recognition. Among them, these detection is the most fundamental and important one.
This paper researches on some popular algorithms inside and outside of our country. Meanwhile, these algorithms are comparably analyzed about their merits and shortcomings. We rise a rapid detection and tracking algorithm based on the color and character of human face. A prototype of the system applicable to the examination of the actual online network authentication. In this paper as an ordinary PC camera image acquisition equipment to collect the data source for streaming video, video streaming interception in the single-frame images, through the conversion of color space, color modeling Face, after handling Face positioning realized through Image mosaics, edge extraction, notable feature points, and the setting of technical positioning of the Face Image feature points. On the basis of this position through the eyes achieved in video streaming in the face tracking algorithm research.
Test results show that we are realizing the face detection and tracking algorithm applied to the face at close range detection, positioning speed, low error rate seized, can be used to test network authentication.
引文
[1]段锦,张悦,崔卓 计算机系统,Vol.,周春光.基于人脸识别技术的访问控制系统,小型微型机 23,2002(12):1472-14750
[2]梁路宏,艾海周,徐光佑等.人脸检测综述,计算机学报,Vol.25 2002(5):
[3]Yoo T.W,Oh I.S.A fast algorithm for tracking human faces based on chromatic histograms.Pattern Recognition Letters 1999,20(10):967-978
[4]Cai J,Goshtasby A.Detecting human faces in color images.Image and Vision Computing,1999(1),18:63-75
[5]Garcia C,Tzidtas G.Face detection using quantized skin color regions merging and wavelet packet analysis.IEEE Trans.on Multimedia,1999,1(3):264-277
[6]Abdel-Mottaleb M,Elgammal A.Face detection in complex environments from color images.In:Proc.IEEE Conf.on Image Processing,Kobe,Japan,1999,Vol.3(11-13):622-626
[7]Wei G,Sethi I K.Face detection for image annotation.Pattern Recognition Letters,20(1999):1313-1321
[8]Karlekar J,Desai U B.Finding faces in color images using wavelet transform.In:Proc.IEEE Conf.on Image Analysis and Processing,Venice,Italy,1999,1085-1088
[9]Wu H Y,Chen Q,Yachida M.A fuzzy-theory-based face detector.In:Proc.Conf.on Pattern Recognition,Vienna,Austria,1996,Vol.3:406-410
[10]Govindaraju V,Srihari S N,Sher D B.A computational model for face location.In:Proc.IEEE Conf.on Computer Vision,Osaka,Japan,1990,718-721
[11]Yang M H,Ahuja N.Detecting human faces in color images.In Proc.IEEE Conf.on Image Processing,Chicago,1998,
[12]卢春雨,张长水,闻方,阎平凡,基于区域特征的快速人脸检测法,清华大学学报(自然科学版),1999,39(1):101-105
[13]Miao J.,Yin B.C.,Wang K.Q.,et al.A hierachical multiscale and multiangle system for human face detection in a complex background using gravity-center template.Pattern Recognition,1999,32(10):1237-1248
[14]Moghaddam B,Pentland A.Probabilistic visual learning for object representation.IEEE Trans.Pattern Analysis and Machine Intelligence,1997,19(7),696-710
[15]Schneiderman H,Kanade T.Probabilistic modeling of local appearance and spatial relationships for object recognition.In:Proc.IEEE Conf.on Computer Vision and Pattern Recognition,Santa Barbara,California,1998,45-51
[16]Osuna E,Freund R,Girosi F.Training support vector machines:an application to face detection.In:Proc.of Computer Vision and Pattern Recognition,Puerto Rico,1997.130-136
[17]Viola Paul,Jones Michael.Rapid Object Detection using a Boosted Cascade of Simple Features.In:Proc.IEEE Conf.on Computer Vision and Pattern Recognition,Kauai,Hawaii,USA,2001.
[18]张宏林 Visual C++数字图像模式识别技术及工程程实践.北京:人民邮电出版
[19]北京嘉恒中自图像技术有限公司,OK卡用户手册,第十版,北京:北京嘉恒中自图像技术有限公司,2003,06
[20]易丛琴,李丽,孟传良.基于肤色分割的彩色图像的人脸检测 贵州工业大学学报(自然科学版),2003,32(4):6568
[21]B.Froba,C.Kublbeck.Robust Face Detection at Video Frame Rate Based on Edge Orientation Features.Proc.Fifth IEEE AFGR.2002:327-332
[22]J.Miao,B.Yin,K.Wang,and L.Shen,et al.A Hierarchical Multiscale and Multiangle System for Human Face Detection in a Complex Background Using Gravity-Center Template.Pattern Recognition.1999(32,7):1237-1248
[23]陶霖密,彭林,徐光佑.人体的肤色特征.软件学报,2001,12(7):1032-1041
[24]周心明,兰赛,徐燕.图象处理中几种边缘检测算法的比较,现代电力,Vol.17,No.3,2000,8:65 69
[25]林锋,刘政凯等.基于图象分割的人眼边缘提取方法研究,计算机应用研究,2000,8:100-1020
[26]汪孔桥.面向头肩序列图象的人脸检测、跟踪与编码技术的研究.中国科技大学博士论文.1999
[27]梁路宏,艾海舟.基于人脸检测的人脸跟踪算法.计算机工程与应用.2001,37(17):42-45
[28]Giachetti A,Campani M,Torre V.The use of optical flow for road navigation[J].IEEE Transactions on Robotics and Automation,1998,14(1):34-48
[29]Batavia PH,Pomerleau DE,Thorpe CE,Overtaking vehicle detection using implicit optical flow[C].In:proc.of th IEEE Conference on Intelligent Transportation System.Boston,Massachusetts,1997,729-734
[30]S.M Smith,IM Brady.ASSEKT-2:Real-time motion segmentation and shape tracking[J].EE.rans.on Pattern Analysis and Machine Intelligence,1995,17814-820
[31]王志良,韩忠涛,刘冀伟,韩鸿哲.基于背景差的实时行人检测和行人记数系统.第二届全国 智能视觉监控学术会议.北京,2003:82-86
[32]McKenna S et al.Tracking groups of people[J].Computer Vision and Image Understanding,2000,81-82
[33]Stauffer C,Grimson W.Adaptive background mixture models for real-time tracking[C].In:Proc IEEE Conference on Computer Vision and Patter Recognition,Fort Collims,Colorado,1999,2:246-252
[2]梁路宏,艾海周,徐光佑等.人脸检测综述,计算机学报,Vol.25 2002(5):
[3]Yoo T.W,Oh I.S.A fast algorithm for tracking human faces based on chromatic histograms.Pattern Recognition Letters 1999,20(10):967-978
[4]Cai J,Goshtasby A.Detecting human faces in color images.Image and Vision Computing,1999(1),18:63-75
[5]Garcia C,Tzidtas G.Face detection using quantized skin color regions merging and wavelet packet analysis.IEEE Trans.on Multimedia,1999,1(3):264-277
[6]Abdel-Mottaleb M,Elgammal A.Face detection in complex environments from color images.In:Proc.IEEE Conf.on Image Processing,Kobe,Japan,1999,Vol.3(11-13):622-626
[7]Wei G,Sethi I K.Face detection for image annotation.Pattern Recognition Letters,20(1999):1313-1321
[8]Karlekar J,Desai U B.Finding faces in color images using wavelet transform.In:Proc.IEEE Conf.on Image Analysis and Processing,Venice,Italy,1999,1085-1088
[9]Wu H Y,Chen Q,Yachida M.A fuzzy-theory-based face detector.In:Proc.Conf.on Pattern Recognition,Vienna,Austria,1996,Vol.3:406-410
[10]Govindaraju V,Srihari S N,Sher D B.A computational model for face location.In:Proc.IEEE Conf.on Computer Vision,Osaka,Japan,1990,718-721
[11]Yang M H,Ahuja N.Detecting human faces in color images.In Proc.IEEE Conf.on Image Processing,Chicago,1998,
[12]卢春雨,张长水,闻方,阎平凡,基于区域特征的快速人脸检测法,清华大学学报(自然科学版),1999,39(1):101-105
[13]Miao J.,Yin B.C.,Wang K.Q.,et al.A hierachical multiscale and multiangle system for human face detection in a complex background using gravity-center template.Pattern Recognition,1999,32(10):1237-1248
[14]Moghaddam B,Pentland A.Probabilistic visual learning for object representation.IEEE Trans.Pattern Analysis and Machine Intelligence,1997,19(7),696-710
[15]Schneiderman H,Kanade T.Probabilistic modeling of local appearance and spatial relationships for object recognition.In:Proc.IEEE Conf.on Computer Vision and Pattern Recognition,Santa Barbara,California,1998,45-51
[16]Osuna E,Freund R,Girosi F.Training support vector machines:an application to face detection.In:Proc.of Computer Vision and Pattern Recognition,Puerto Rico,1997.130-136
[17]Viola Paul,Jones Michael.Rapid Object Detection using a Boosted Cascade of Simple Features.In:Proc.IEEE Conf.on Computer Vision and Pattern Recognition,Kauai,Hawaii,USA,2001.
[18]张宏林 Visual C++数字图像模式识别技术及工程程实践.北京:人民邮电出版
[19]北京嘉恒中自图像技术有限公司,OK卡用户手册,第十版,北京:北京嘉恒中自图像技术有限公司,2003,06
[20]易丛琴,李丽,孟传良.基于肤色分割的彩色图像的人脸检测 贵州工业大学学报(自然科学版),2003,32(4):6568
[21]B.Froba,C.Kublbeck.Robust Face Detection at Video Frame Rate Based on Edge Orientation Features.Proc.Fifth IEEE AFGR.2002:327-332
[22]J.Miao,B.Yin,K.Wang,and L.Shen,et al.A Hierarchical Multiscale and Multiangle System for Human Face Detection in a Complex Background Using Gravity-Center Template.Pattern Recognition.1999(32,7):1237-1248
[23]陶霖密,彭林,徐光佑.人体的肤色特征.软件学报,2001,12(7):1032-1041
[24]周心明,兰赛,徐燕.图象处理中几种边缘检测算法的比较,现代电力,Vol.17,No.3,2000,8:65 69
[25]林锋,刘政凯等.基于图象分割的人眼边缘提取方法研究,计算机应用研究,2000,8:100-1020
[26]汪孔桥.面向头肩序列图象的人脸检测、跟踪与编码技术的研究.中国科技大学博士论文.1999
[27]梁路宏,艾海舟.基于人脸检测的人脸跟踪算法.计算机工程与应用.2001,37(17):42-45
[28]Giachetti A,Campani M,Torre V.The use of optical flow for road navigation[J].IEEE Transactions on Robotics and Automation,1998,14(1):34-48
[29]Batavia PH,Pomerleau DE,Thorpe CE,Overtaking vehicle detection using implicit optical flow[C].In:proc.of th IEEE Conference on Intelligent Transportation System.Boston,Massachusetts,1997,729-734
[30]S.M Smith,IM Brady.ASSEKT-2:Real-time motion segmentation and shape tracking[J].EE.rans.on Pattern Analysis and Machine Intelligence,1995,17814-820
[31]王志良,韩忠涛,刘冀伟,韩鸿哲.基于背景差的实时行人检测和行人记数系统.第二届全国 智能视觉监控学术会议.北京,2003:82-86
[32]McKenna S et al.Tracking groups of people[J].Computer Vision and Image Understanding,2000,81-82
[33]Stauffer C,Grimson W.Adaptive background mixture models for real-time tracking[C].In:Proc IEEE Conference on Computer Vision and Patter Recognition,Fort Collims,Colorado,1999,2:246-252