未来智讯 > 人脸语音识别论文 > 基于核非负稀疏表示的人脸识别

基于核非负稀疏表示的人脸识别

发布时间:2017-12-07 09:10:00 文章来源:未来智讯    
    关键词:人脸识别;稀疏表示;核函数;局部二元特征;汉明核
    中图分类号: TP391.413
    文献标志码:A
    Abstract: Face recognition is one of important topics in computer vision and pattern recognition. A novel kernelbased nonnegative sparse representation (KNSR) method was presented based on this topic and was tested on face databasesfor face recognition. The contributions were mainly three aspects: First, the nonnegative constraints on representation coefficients were introduced into the Sparse Representation (SR) and the kernel function was exploited to depict nonlinear relationships among different samples, based on which the corresponding objective function was proposed. Second, a multiplicative gradient descent method was proposed to solve the proposed objective function, which could achieve the global optimum value in theory. Finally, local binary feature and the Hamming kernel were used to model the nonlinear relationships among face samples and therefore achieved robust face recognition. The experimental results on some challenging face databases demonstrate that the proposed algorithm has higher recognition rates in comparison with algorithms of Nearest Neighbor (NN), Support Vector Machine (SVM), Nearest Subspace (NS), SR and Collaborative Representation (CR), and achieves about 99% recognition rates on both YaleB and AR databases.
    Key words: face recognition; sparse representation; kernel function; local binary feature; Hamming kernel
    0引言
    人脸识别是计算机视觉和模式识别领域的一个重要研究课题,并且在安防监控、人机交互、增强现实等方面有着广泛的应用。人脸识别算法已经被研究了很多年,很多学者提出了不同的人脸识别算法(如基于主成分分析的人脸识别算法[1]、基于线性判决分析的人脸识别算法[2]、基于局部保留映射的人脸识别算法[3]、弹性束图匹配算法[4]、基于局部二值特征的人脸识别算法[5]、基于Gabor相位直方图的人脸识别算法[6]、最近子空间算法[7]等),但是由于人脸图像外观存在大量内在和外在的变化(如年龄、表情、光照、遮挡等),人脸识别依然是一个极富有挑战性的研究课题。
    近年来,随着稀疏表示和压缩感知理论的发展和完善,它们被越来越多用于处理计算机视觉和模式识别问题(如人脸识别[8]、图像超分辨[9]、图像修复[10]、视觉跟踪[11]等方面)。2009年,Wright等[8]提出了稀疏表示分类框架,并将其用于人脸识别。在该算法框架中,一个待测试的人脸图像可以看作所有训练人脸图像的稀疏线性组合,并利用L1最小化技术来求解组合系数(也称为表示系数)。在获得表示系数后,利用最小类残留(即类重构误差最小)的方式进行分类。受到该算法的启发,很多学者开始尝试利用稀疏表示算法来进行人脸识别。Yang等[12]在稀疏表示分类框架中利用Gabor特征取代裸像素特征,并通过学习一个遮挡字典来处理遮挡的人脸图像。Zhou等[13]稀疏表示分类框架中引入了遮挡像素的空间连续性,利用一阶马尔可夫随机场来建模连续的遮挡。文献[14-15]提出了基于核的稀疏表示分类算法,并将其用于人脸识别和分类领域。此外,Zhang等[16]还对文献[8]中的稀疏表示分类框架进行了深入的分析,指出在稀疏分类框架中协同表示的作用大于稀疏限制的作用,据此提出了一种基于均方L2范数限制的协同表示算法,很大程度上提升了识别的速度。
         本文提出了一种基于核非负稀疏表示的人脸识别算法。首先,该算法在稀疏表示的基础上引入了对表示系数的非负限制,保证了人脸图像数据的加性组合,并通过利用核函数进行隐式的非线性映射,使得分类框架能够利用非线性特征或度量;其次,本文对目标函数进行了变型和松弛,将其转化为一个非负最小均方问题,并设计了乘性梯度下降算法进行求解,理论上可以保证获得全局最优解;最后,本文实现了利用核非负稀疏表示进行人脸识别实验,实验利用了局部二元特征和汉明核来捕捉人脸数据间的非线性关系。实验结果表明,与其他算法相比,本文提出的算法具有很好的识别精度。
    3.3在AR数据库上的实验结果
    AR人脸数据库[21]包含来自126个人超过4000幅正面人脸图像。参考文献[8],本文选择了50个男性的人脸图像和50个女性的人脸图像(共计100个人)。每一个人包含14张无遮挡的图像、6张被眼镜遮挡的图像和6张被围巾遮挡的图像(图2为AR遮挡数据库图像示意)。所有的人脸图像均被归一化到32×32大小。为了评估算法在无遮挡情况下的表示,在无遮挡数据子集上对每个人随机选择l幅图像组成训练集用于训练,其余的无遮挡图像用作测试(l=2,4,6,8,10,12)。表2显示了AR数据库上无遮挡人脸识别的测试结果。此外,本文来利用AR数据库评测算法在处理实际遮挡或伪装时的表现(此时利用无遮挡的人脸数据作为训练样本,遮挡的人脸数据作为测试样本),其中粗体代表最好的识别结果。表3显示了在遮挡数据库子集(眼镜遮挡和围巾遮挡)不同人脸识别算法的结果,其中SR(Patch)和CR(Patch)代表文献[8]和[16]中采用分块策略处理遮挡的结果,粗体代表最好的识别结果。从表2~3中可以看出,与其他算法相比,本文提出的KNSR算法(特别在小训练样本和遮挡情况时)具有明显的精度优势。
    4结语
    本文提出了一种新颖的基于核非负稀疏表示的人脸识别算法。首先,本文在稀疏表示的基础上引入了对系数的非负限制,以及利用隐式核映射来处理非线性表示问题,并据此提出了相应的目标函数;其次,本文推导了一种乘性梯度下降准则来求解所提出的目标函数;最后,本文利用局部二元特征和汉明核来建模人脸样本的非线性关系,从而实现鲁棒的人脸识别。实验结果表明在Extend Yale B和AR数据库上本文算法取得了不错的识别结果。
    参考文献:
    [1]TURK M, PENTLAND A. Eigenfaces for recognition [J]. Journal of Cognitive Neuroscience, 1991, 3(1): 71-86.
    [2]ETEMAD K, CHELLAPPA R. Discriminant analysis for recognition of human face images [J]. Journal of the Optical Society of America, 1997, 14(1): 1724-1733.
    [3]HE X, YAN S, HU Y, et al. Face recognition using Laplacianfaces [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(3): 328-340.
    [4]WISKOTT L, FELLOUS J M, KRUGER N, et al. Face recognition by elastic bunch graph matching [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1997, 19(7): 775-779.
    [5]AHONEN T, HADID A, PIETIKAINEN M. Face recognition with local binary patterns [C]// Proceedings of the 8th European Conference on Computer Vision, LNCS 3021. Berlin: Springer, 2004: 469-481.
    [6]ZHANG B, SHAN S, CHEN X, et al. Histogram of Gabor Phase Patterns (HGPP): a novel object representation approach for face recognition [J]. IEEE Transactions on Image Processing, 2007, 16(1): 57-68.
    [7]LEE K C, HO J, KRIEGMAN D J. Acquiring linear subspaces for face recognition under variable lighting [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(5): 684-698.
    [8]WRIGHT J, YANG A Y, GANESH A, et al. Robust face recognition via sparse representation [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(2): 210-227.
    [9]YANG J, WRIGHT J, HUANG T S, et al. Image superresolution via sparse representation [J]. IEEE Transactions on Image Processing, 2010, 19(11): 2861-2873.
         [10]WRIGHT J, MA Y, MARAL J, et al. Sparse representation for computer vision and pattern recognition [J]. Proceedings of the IEEE, 2010, 98(6): 1031-1044.
    [11]WANG D, LU H C, YANG M H. Online object tracking with sparse prototypes [J]. IEEE Transactions on Image Processing, 2013, 22(1): 314-325.
    [12]YANG M, ZHANG L. Gabor feature based sparse representation for face recognition with Gabor occlusion dictionary [C]// Proceedings of the 11th European Conference on Computer Vision, LNCS 6316. Berlin: Springer, 2010: 448-461.
    [13]ZHOU Z, WAGNER A, MOBAHI H, et al. Face recognition with contiguous occlusion using Markov random fields [C]// Proceedings of IEEE 12th International Conference on Computer Vision. Washington, DC: IEEE Computer Society, 2009: 1050-1057.
    [14]KANG C, LIAO S, XIANG S, et al. Kernel sparse representation with local patterns for face recognition [C]// Proceedings of the 18th IEEE International Conference on Image Processing. Washington, DC: IEEE Computer Society, 2011: 3009-3012.
    [15]GAO S, TSANG I W H, CHIA L T. Sparse representation with kernels [J]. IEEE Transactions on Image Processing, 2013, 22(2):423-434.
    [16]ZHANG L, YANG M, FENG X. Sparse representation or collaborative representation: which helps face recognition? [C]// Proceedings of the 2011 International Conference on Computer Vision. Washington, DC: IEEE Computer Society, 2011: 471-478.
    [17]LEE D D, SEUNG H S. Learning the parts of objects by nonnegative matrix factorization [J]. Nature, 1999, 401(6755): 788-791.
    [18]ZHOU H, TREVOR T. Regularization and variable selection via the elastic net [J]. Journal of the Royal Statistical Society, Series B, 2005, 67(2): 301-320.
    [19]OJALA T, PIETIKAINEN M, MAENPAA T. Multiresolution grayscale and rotation invariant texture classification with local binary patterns [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(7): 971-987.
    [20]GEORGHIADES A S, BELHUMEUR P N, KRIEGMAN D J. From few to many: illumination cone models for face recognition under variable lighting and pose [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2001, 23(6): 643-660.
    [21]MARTINEZ A, BENAVENTE R. The AR face database, CVC technical report 24 [R]. West Lafayette: Purdue University, Computer Vision Center, 1998.
转载请注明来源。原文地址:https://www.7428.cn/page/2017/1207/9199/
 与本篇相关的热门内容: