光照补偿在人像识别中的应用及改善
摘要
随着社会的发展，各个方面对快速有效的自动身份验证的要求日益迫切。生物特征是人的内在属性，具有很强的自身稳定性和个体差异性，是身份验证的最理想的依据。这样，“生物特征识别技术”，因其良好的安全性越来越多地应用于身份识别。利用人脸特征进行身份验证又是最自然直接的手段 ,相比其它人体生物特征它具有直接、友好、方便的特点 ,易于为用户接受等优点，成为其中很有前景的一部分。目前人脸检测与识别是计算机视觉和人机交互研究领域的一个重要方向。将这项任务自动化是非常有意义的。计算机技术和人工智能领域研究的发展使得这方面的研究有了长足的进步。同时，根据计算机技术的发展趋势，智能人机界面应该取代传统的键盘鼠标，成为更加安全方便的人机交互接口。这里面最为基础的就是机器应该能够轻易地知道是谁在它的视野中，并知道在计算机可检测范围的具体位置。
人脸识别的研究最近几年再度受到普遍重视，它与指纹识别、视网膜识别等同属于生物特征识别技术范畴 ,在诸如安全检查、保安管理等方面有着十分重要的应用价值。目前对于如何检测人脸或输入人脸的研究是人们普遍关心的问题。在已发现的文章中 ,多数的研究是一般环境下的科学研究，不具备实用的意义，很难实际应用。因为人脸图像的获取过程不同，导致二维图像信息在质量、几何、光线上都有内在的不同，此外还有脸部受到遮挡和化妆等因素的影响。但是，更内在的原因是，人脸是具有高度相似性的非刚体。人脸不同于普通物体，不同人的脸具有高度的相似性，同一人的脸又具有不同的状态，这使得人脸识别问题不同于普通物体的识别问题。目前，许多研究机构致力于这一领域的研究，取得了丰硕的理论成果并有不同的应用软件应运而生。尽管如此，可以说，还没有一个完全的解决方案可以在毫无约束的情况下出色地完成人脸定位，识别任务。
在应用领域上，目前可能的识别任务主要可以归为两类：身份识别/辨认/匹配(Recognition/ Identification/ match)、身份验证/证实/监督(Verification/ Authentification /Surveillance)。前者在应用上的典型实例就是公安刑侦追逃。第二种应用情形的典型实例是身份证件的鉴别、自动门禁控制系统、公司部门的人员考勤、家庭安全等领域。目前，这一领域的产品正在受到越来越多的关注。
环境光线的变化是影响人像识别精度的主要因素之一。实验室环境下的实现方法，可能在变化的环境光线下，变得质量下降或是不可用。本论文从
发现光照恒常性规律的角度入手，对几种补偿光照、或改善光照的传统方法进行了讨论，并提出了一种基于小波的可变光照改善方法。实验结果证明了此种方法的有效性，并可以推广到实际的应用系统中，提高系统鲁棒性和适应性。
本文对光照补偿和改善做了研究，发现只有在规定了姿势和光照的条件下，特定的光照补偿算法对人脸的检测和识别才是可靠的。实验中可知，在对图像进行三级以上小波变换后，所得到的近似分量已经不包含原图像的基本特征，而只反映图像的明暗程度，所以，在多级小波变换的基础上对图像进行光照改善有很好的效果，本文所做的工作正是基于这种理论而得出的。
本文在已有的理论研究成果的基础上做了进一步的研究和实验工作，并将这些工作部分转化成软件产品，取得了一定的应用。
一个自动人脸识别系统的主要组成部分是图像获取、人脸的检测定位、人脸图像预处理、特征提取、分类器设计和决策部分。较为重要和困难的问题集中在人脸检测定位，特征提取，分类器设计上。
在上述的理论和实验的基础之上，我们研制了可以用于实际生产生活的软件产品――自动人像识别门禁系统，人像屏幕保护和人像考勤系统。并将之应用到边防局，银行，公安等部门。这些是身份验证类的应用，来访者手持磁卡，系统可以得知人员的ID，通过图像判断其是否有权限通过。
在应用中研究了不同的硬件，并结合自身的应用开发了单机和网络版的产品，增加了产品应用的广泛性。
由于时间的关系，当前的算法对环境变化的适应度还不够好；开发的软件只能应用于特定的环境；距离真正的广泛的商业化还有很大差距，这些都是值得今后继续研究的问题。
Biometric identification systems, which use physical features to check a person's identity, ensure much greater security than password and number systems. Identifying a human individual from his or her face is one of the most non-intrusive modalities in biometrics. The capability of finding and recognizing a face in a random scene is important in everyone’s activities. So it’s very significant that automate this task. Development in computer technology and artificial intelligence make research on these fields be possible. At the same time, it also spurs the efforts on intelligent interface between human and machine. It is fundamental that computer should know who are in its eyeshot.
However, it is also one of the most challenging problems. At first, the face images are obtained by different way or in different condition, so they have substantial difference in quality, geometry, illumination, etc. In addition, it also exits the makeup and face-painting influence. But the most essential reason is that face is a kind of non-rigid object that has highly similarity. Different person’s faces have similar shape and structure, and one person’s face has different state. In the past decade, many research groups make great efforts on it and a series of successes have made general personal identification appear not only technically feasible but also economically practical. However, no perfect solution can accomplish this task under the non-constraint condition.
It is of particular interest in a wide variety of applications. In fact, face recognition technology has two kinds of application: recognition/ identification/ match and verification/ authentification/ surveillance. A typical example about the former one is applications in law enforcement for mug-shot identification. The latter application is also broad, such as verification for personal identification, gateways to limited access areas, authentification for ATM and family security, etc. In brief, one is concluding who is he/she and the other is deciding is he/she somebody.
The illumination fluctuation is one of the important factors that
influence the precision of face recognition. When environmental illumination is changed, the implementation method in laboratory maybe becomes unavailability. Several traditional methods is discussed for compensation and improving Variable illumination . We want to find the invariance of illumination. Then a novel algorithm based on Wavelet Analysis is presented in this paper. Several empirical tests are given to demonstrate the effectiveness and quality of our method. The method can be applied to a real system , can raise system’s robusticity and adaptability.
This text compensate and improve and make and study to illumination, Find only under fixed posture, specific illumination compensate between algorithm and people the measuring and it discerns to be reliable s of face. Experiment know, in carrying on and after tertiary little waveseses vary, have to go to and similar to weight include essential feature who art of work look like already to picture, So, carry on the illumination and improve very fine result to the picture on the basis of varying in many grades of little unexpected turn oves, Work that this text cook draw because of these kind of theory.It reflects the light and shade degreeses of picture only however, carry on illumination improve and it has to be very kind result to picture on the basis of varying in many grades of little unexpected turn oves, The work made of this text was just drawn because of this kind of theory.
This paper is interested in doing some father work in theory research and practical experiment. It also outlined two instances concerned.
A typical automatic face recognition system consists of the following function: image obtaining, face detection and location, image preprocessing, feature extraction, classifier designing and deciding. Most research focuses on face detection, feature extraction and classifier designing.
Based on the above theory and experiment, software has been d
人脸识别的研究最近几年再度受到普遍重视，它与指纹识别、视网膜识别等同属于生物特征识别技术范畴 ,在诸如安全检查、保安管理等方面有着十分重要的应用价值。目前对于如何检测人脸或输入人脸的研究是人们普遍关心的问题。在已发现的文章中 ,多数的研究是一般环境下的科学研究，不具备实用的意义，很难实际应用。因为人脸图像的获取过程不同，导致二维图像信息在质量、几何、光线上都有内在的不同，此外还有脸部受到遮挡和化妆等因素的影响。但是，更内在的原因是，人脸是具有高度相似性的非刚体。人脸不同于普通物体，不同人的脸具有高度的相似性，同一人的脸又具有不同的状态，这使得人脸识别问题不同于普通物体的识别问题。目前，许多研究机构致力于这一领域的研究，取得了丰硕的理论成果并有不同的应用软件应运而生。尽管如此，可以说，还没有一个完全的解决方案可以在毫无约束的情况下出色地完成人脸定位，识别任务。
在应用领域上，目前可能的识别任务主要可以归为两类：身份识别/辨认/匹配(Recognition/ Identification/ match)、身份验证/证实/监督(Verification/ Authentification /Surveillance)。前者在应用上的典型实例就是公安刑侦追逃。第二种应用情形的典型实例是身份证件的鉴别、自动门禁控制系统、公司部门的人员考勤、家庭安全等领域。目前，这一领域的产品正在受到越来越多的关注。
环境光线的变化是影响人像识别精度的主要因素之一。实验室环境下的实现方法，可能在变化的环境光线下，变得质量下降或是不可用。本论文从
发现光照恒常性规律的角度入手，对几种补偿光照、或改善光照的传统方法进行了讨论，并提出了一种基于小波的可变光照改善方法。实验结果证明了此种方法的有效性，并可以推广到实际的应用系统中，提高系统鲁棒性和适应性。
本文对光照补偿和改善做了研究，发现只有在规定了姿势和光照的条件下，特定的光照补偿算法对人脸的检测和识别才是可靠的。实验中可知，在对图像进行三级以上小波变换后，所得到的近似分量已经不包含原图像的基本特征，而只反映图像的明暗程度，所以，在多级小波变换的基础上对图像进行光照改善有很好的效果，本文所做的工作正是基于这种理论而得出的。
本文在已有的理论研究成果的基础上做了进一步的研究和实验工作，并将这些工作部分转化成软件产品，取得了一定的应用。
一个自动人脸识别系统的主要组成部分是图像获取、人脸的检测定位、人脸图像预处理、特征提取、分类器设计和决策部分。较为重要和困难的问题集中在人脸检测定位，特征提取，分类器设计上。
在上述的理论和实验的基础之上，我们研制了可以用于实际生产生活的软件产品――自动人像识别门禁系统，人像屏幕保护和人像考勤系统。并将之应用到边防局，银行，公安等部门。这些是身份验证类的应用，来访者手持磁卡，系统可以得知人员的ID，通过图像判断其是否有权限通过。
在应用中研究了不同的硬件，并结合自身的应用开发了单机和网络版的产品，增加了产品应用的广泛性。
由于时间的关系，当前的算法对环境变化的适应度还不够好；开发的软件只能应用于特定的环境；距离真正的广泛的商业化还有很大差距，这些都是值得今后继续研究的问题。
Biometric identification systems, which use physical features to check a person's identity, ensure much greater security than password and number systems. Identifying a human individual from his or her face is one of the most non-intrusive modalities in biometrics. The capability of finding and recognizing a face in a random scene is important in everyone’s activities. So it’s very significant that automate this task. Development in computer technology and artificial intelligence make research on these fields be possible. At the same time, it also spurs the efforts on intelligent interface between human and machine. It is fundamental that computer should know who are in its eyeshot.
However, it is also one of the most challenging problems. At first, the face images are obtained by different way or in different condition, so they have substantial difference in quality, geometry, illumination, etc. In addition, it also exits the makeup and face-painting influence. But the most essential reason is that face is a kind of non-rigid object that has highly similarity. Different person’s faces have similar shape and structure, and one person’s face has different state. In the past decade, many research groups make great efforts on it and a series of successes have made general personal identification appear not only technically feasible but also economically practical. However, no perfect solution can accomplish this task under the non-constraint condition.
It is of particular interest in a wide variety of applications. In fact, face recognition technology has two kinds of application: recognition/ identification/ match and verification/ authentification/ surveillance. A typical example about the former one is applications in law enforcement for mug-shot identification. The latter application is also broad, such as verification for personal identification, gateways to limited access areas, authentification for ATM and family security, etc. In brief, one is concluding who is he/she and the other is deciding is he/she somebody.
The illumination fluctuation is one of the important factors that
influence the precision of face recognition. When environmental illumination is changed, the implementation method in laboratory maybe becomes unavailability. Several traditional methods is discussed for compensation and improving Variable illumination . We want to find the invariance of illumination. Then a novel algorithm based on Wavelet Analysis is presented in this paper. Several empirical tests are given to demonstrate the effectiveness and quality of our method. The method can be applied to a real system , can raise system’s robusticity and adaptability.
This text compensate and improve and make and study to illumination, Find only under fixed posture, specific illumination compensate between algorithm and people the measuring and it discerns to be reliable s of face. Experiment know, in carrying on and after tertiary little waveseses vary, have to go to and similar to weight include essential feature who art of work look like already to picture, So, carry on the illumination and improve very fine result to the picture on the basis of varying in many grades of little unexpected turn oves, Work that this text cook draw because of these kind of theory.It reflects the light and shade degreeses of picture only however, carry on illumination improve and it has to be very kind result to picture on the basis of varying in many grades of little unexpected turn oves, The work made of this text was just drawn because of this kind of theory.
This paper is interested in doing some father work in theory research and practical experiment. It also outlined two instances concerned.
A typical automatic face recognition system consists of the following function: image obtaining, face detection and location, image preprocessing, feature extraction, classifier designing and deciding. Most research focuses on face detection, feature extraction and classifier designing.
Based on the above theory and experiment, software has been d
引文
[1] Alex(Sandy) Pentland,Tanzeem Choudhury. "Face Recognition for Smart Environments". IEEE Computer ,February 2000 ,p50-p55
[2] 志华，皇甫杰，张宏江，陈祖翰，“基于神经网络集成的多视角人脸识别”，计算机研究与发展，Vol.38,No.10,2001.10
[3] 彭进业，王大凯，俞卞章，李楠，“基于小波分解系数的贝叶斯人脸识别方法”，光子学报，Vol.30,No.10,2001.10
[4] 王蕴红，谭铁牛，朱勇，"基于奇异值分解和数据融合的脸像鉴别",计算机学报，Vol.23, No.6, 2000.6
[5] Chellapps R. Wilson C L, Sirohey S. Human and machine recognition of faces : A survey, Proc. Of The Tree, 1995,83(5): 705－740
[6] Athinodoros S. Georghiades, Peter N. Bellhumeur, David J. Kriegman. From Few to Many: Generative Models for Recognition Under Variable Pose and Illumination[J]. IEEE Trans. Pattern Anal. Mach. Intelligence, 2001:643-660.
[7] Henry A. Rowley, Shumeet Baluja, Takeo Kanade. Neural Network-Based Face Detection[A].? IEEE PAMI[C],January 1998.
[8] Warrington E, Taylor A. Two Cstegorical Stages of Object Recognition [J]. Perception, 1978, 7: 152-164.
[9] Govindaraju V, Sher D B, Srihari R, Srihari S N. Locating Human Faces in Newspaper Photograph [A]. Proc CVPP89 [C], 1989. 549-554.
[10] Miao J, Yia B C, Wang K Q, et al. A hierarchical multiscale and multiangle system for human face detection in a complex background using gravity-center template [J]. Pattern Recognition, 1999, 32(7): 1237-1248.
[11] Buhmann J, Lades M, Beckman F. A Silicon Retina Object Recognition [R]. Techaical Report No 8596-CS, Institut fur Informatik, Universitat Bonn, 1993.
[12] Edeiman S, Reisfeld D, Yeshurun Y. A System for Face Recognition that Learns from Examples [A]. Proc, European Conf Computer Vision, Sandini S, ed [C]. Springer-Verlag, 1992. 787-791.
[13] Reisfeld D, Yeshurun Y. Robust Detection of Facial Features by Generalized Symmetry [A]. ICPR A, 1992, 117-120.
[14] Miao J, Zhang H, Gao W, Liu H, FaceTracher: A Human Face Tracking and Facial Organ Localizing System [A]. Proc the 8th International Conference on
Computer Vision [C], 2001.
[15] Mallat S, A theory of Multiresolution signal decomposition: the wavelet representation. IEEE Tran. PAMI. 1989, 11: 674-693
[16] Thomas Fromherz, Peter Stucki, Martin Bichsel. A Survey of Face Recognition[R]. MML Technical Report,1997,No. 9701.
[17]Ming-Hsuan Yang, Narendra Ahuja, and David Kriegman. A survey on Face Detection Methods[A]. DRAFF[C] March 6,1999.
[18] Motonori Doi, Kosuke Sato and Kunihiro Chihara. A Robust Face Identification against Lighting Fluctuation for Lock Control[A]. Proceedings of IEEE International Conference on Automatic Face and Gesture Recognition[C] ,1998:42-47 .
[19] Hany Farid and Edward H. Adelson. Separating Reflections and Lighting Using Independent Components Analysis[A]. In Proceedings of IEEE Computer Society Conference On Computer Vision and Pattern Recognition(CVPR)[C] ,1999, 1: 262-267.
[20] 边肇祺，张学工等编著，《模式识别》，清华大学出版社，2000年，第二版，p223－p227
[21] Moses Y. Ullman S. Limitation of Non-Model-Based Recognition Schemes [A]. Proc ECCV-92, Sandini G, ed [C]. Springer-Verlag, 1992. 820-828
[22] Mallat S, Multiresolution approximations and orthonormal bases of L2(R). Trans. Amer. Math. Soc. 1989, 315:69-87
[23] Mallat S, Multiresolution representation and wavelet. Ph. D. Thesis. Univ. of Pennsylvania, Philadelphia, 1988, 41(7): 909-996
[24] Chen Dongqing, Xie Hongbo, Yu Daoyin. Enhancement of Endo-scopic Image Based on the Illumination-Reflectance Model [J].. JIG, 5: 446-450
[2] 志华，皇甫杰，张宏江，陈祖翰，“基于神经网络集成的多视角人脸识别”，计算机研究与发展，Vol.38,No.10,2001.10
[3] 彭进业，王大凯，俞卞章，李楠，“基于小波分解系数的贝叶斯人脸识别方法”，光子学报，Vol.30,No.10,2001.10
[4] 王蕴红，谭铁牛，朱勇，"基于奇异值分解和数据融合的脸像鉴别",计算机学报，Vol.23, No.6, 2000.6
[5] Chellapps R. Wilson C L, Sirohey S. Human and machine recognition of faces : A survey, Proc. Of The Tree, 1995,83(5): 705－740
[6] Athinodoros S. Georghiades, Peter N. Bellhumeur, David J. Kriegman. From Few to Many: Generative Models for Recognition Under Variable Pose and Illumination[J]. IEEE Trans. Pattern Anal. Mach. Intelligence, 2001:643-660.
[7] Henry A. Rowley, Shumeet Baluja, Takeo Kanade. Neural Network-Based Face Detection[A].? IEEE PAMI[C],January 1998.
[8] Warrington E, Taylor A. Two Cstegorical Stages of Object Recognition [J]. Perception, 1978, 7: 152-164.
[9] Govindaraju V, Sher D B, Srihari R, Srihari S N. Locating Human Faces in Newspaper Photograph [A]. Proc CVPP89 [C], 1989. 549-554.
[10] Miao J, Yia B C, Wang K Q, et al. A hierarchical multiscale and multiangle system for human face detection in a complex background using gravity-center template [J]. Pattern Recognition, 1999, 32(7): 1237-1248.
[11] Buhmann J, Lades M, Beckman F. A Silicon Retina Object Recognition [R]. Techaical Report No 8596-CS, Institut fur Informatik, Universitat Bonn, 1993.
[12] Edeiman S, Reisfeld D, Yeshurun Y. A System for Face Recognition that Learns from Examples [A]. Proc, European Conf Computer Vision, Sandini S, ed [C]. Springer-Verlag, 1992. 787-791.
[13] Reisfeld D, Yeshurun Y. Robust Detection of Facial Features by Generalized Symmetry [A]. ICPR A, 1992, 117-120.
[14] Miao J, Zhang H, Gao W, Liu H, FaceTracher: A Human Face Tracking and Facial Organ Localizing System [A]. Proc the 8th International Conference on
Computer Vision [C], 2001.
[15] Mallat S, A theory of Multiresolution signal decomposition: the wavelet representation. IEEE Tran. PAMI. 1989, 11: 674-693
[16] Thomas Fromherz, Peter Stucki, Martin Bichsel. A Survey of Face Recognition[R]. MML Technical Report,1997,No. 9701.
[17]Ming-Hsuan Yang, Narendra Ahuja, and David Kriegman. A survey on Face Detection Methods[A]. DRAFF[C] March 6,1999.
[18] Motonori Doi, Kosuke Sato and Kunihiro Chihara. A Robust Face Identification against Lighting Fluctuation for Lock Control[A]. Proceedings of IEEE International Conference on Automatic Face and Gesture Recognition[C] ,1998:42-47 .
[19] Hany Farid and Edward H. Adelson. Separating Reflections and Lighting Using Independent Components Analysis[A]. In Proceedings of IEEE Computer Society Conference On Computer Vision and Pattern Recognition(CVPR)[C] ,1999, 1: 262-267.
[20] 边肇祺，张学工等编著，《模式识别》，清华大学出版社，2000年，第二版，p223－p227
[21] Moses Y. Ullman S. Limitation of Non-Model-Based Recognition Schemes [A]. Proc ECCV-92, Sandini G, ed [C]. Springer-Verlag, 1992. 820-828
[22] Mallat S, Multiresolution approximations and orthonormal bases of L2(R). Trans. Amer. Math. Soc. 1989, 315:69-87
[23] Mallat S, Multiresolution representation and wavelet. Ph. D. Thesis. Univ. of Pennsylvania, Philadelphia, 1988, 41(7): 909-996
[24] Chen Dongqing, Xie Hongbo, Yu Daoyin. Enhancement of Endo-scopic Image Based on the Illumination-Reflectance Model [J].. JIG, 5: 446-450