计算机科学 ›› 2014, Vol. 41 ›› Issue (8): 311-315.doi: 10.11896/j.issn.1002-137X.2014.08.066
王娟,杜海顺,侯彦东,金勇
WANG Juan,DU Hai-shun,HOU Yan-dong and JIN Yong
摘要： 针对投影非负矩阵分解(PNMF)不能揭示数据流形几何结构和判别信息的问题,提出了一种图嵌入投影非负矩阵分解(GEPNMF)特征提取方法。首先分别构建了描述数据流形几何结构和类间分离度的近邻图,然后采用它们的拉普拉斯矩阵设计了一个图嵌入正则项,并将其与PNMF的目标函数融合,以构造GEPNMF的目标函数。在GEPNMF目标函数中引入了图嵌入正则项,使求得的子空间能够在保持数据流形几何结构的同时,类间间距也最大。另外,还在目标函数中引入了一个正交正则项,以确保GEPNMF子空间基向量具有数据局部表示能力。对求解GEPNMF目标函数的累乘更新规则(MUR)进行了详细的推导。在Yale和CMU PIE人脸数据库上的实验结果表明,提出的图嵌入投影非负矩阵分解特征提取方法比PNMF更适用于解决分类问题。
| [1] Lee D D,Seung H S．Learning the parts of objects by non-negative matrix factorization[J]．Nature,1999,401(6755):788-791 [2] 赵海峰,李其超,罗斌,等.基于局部Walsh变换和非负矩阵分解的脑白质图像分割[J]．光电子·激光,2012,3(7):1425-1430 [3] Li S Z,Hou X,Zhang H,et al．Learning spatially localized,parts-based representation[C]∥Proceedings of IEEE International conference on Computer Vision and Pattern Recognition．2001:207-212 [4] Hoyer P O．Non-negative matrix factorization with sparsenessconstraints[J]．Journal of Machine Learning Research,2004(5):1457-1469 [5] Cai D,He X F,Han J W,et al．Graph Regularized Non-negative matrix factorization for Data Representation[J]．IEEE Trans．on Pattern Analysis and Machine Intellligence,2011,33(8):1548-1560 [6] Zhi R C,Flierl M,Ruan Q Q,et al．Graph-preserving sparsenonnegative matrix factorization with application to facial expression recognition[J]．IEEE Trans．on Systems,Man,and Cybernetics-PART B:Cyberneticas,2011,41(1):38-52 [7] Zafeiriou S,Tefas A,Buciu I,et al．Expoiting discriminant information in nonnegative matrix factorization with application to frontal face verification [J]．IEEE Trans．on Neural Networks,2006,17(3):683-695 [8] Yang J,Yan S,Fu Y, et al．Non-negative graph embedding [C]∥Proceedings of IEEE International conference on Computer Vision and Pattern Recognition．2008:1-8 [9] Yan S,Xu D,Zhang B,et al．Graph embedding and extensions:a general framework for dimensionality reduction [J]．IEEE Trans．on Pattern analysis and machine intelligence,2007,29(1):40-51 [10] Guan N,Tao D,Luo Z,et al．Manifold regularized discriminative nonnegative matrix factorization with fast gradient descent [J]．IEEE Trans．on Image Precessing,2011,0(7):2030-2048 [11] Chen H T,Chang H W,Liu T L．Local discriminant embedding and its variants [J]．Proceedings of IEEE International Conference on Computer Vision and Patern Recognition,2005(2):846-853 [12] Ding C,Li T,Jordan M I．Convex and semi-nonnegative matrix factorizations [J]．IEEE Trans．on Pattern Analysis and Machine Intelligence,2010,32(1):45-55 [13] 方蔚涛,马鹏,成正斌,等.二维投影非负矩阵分解算法及其在人脸识别中的应用[J]．自动化学报,2012,8(9):1503-1512 [14] 刘雪松,王斌,张立明.基于非负矩阵分解的高光谱遥感图像混合像元分解[J]．红外与毫米波学报,2011,0(1):1-5 [15] 于红芸,姜涛,关键.SAR图像目标检测的互信息非负矩阵分解算法[J]．中国图象图形学报,2011,6(1):129-134 [16] Yang Z,Oja E．Linear and nonlinear projective nonnegative matrix factorization [J]．IEEE Trans．on Neural Networks,2010,1(5):734-749 [17] Lee D D,Seung H S．Algorithms for non-negative matrxi factori-zation [J]．Advances in Neural Information Processing Systems,MIT Press,2001,3:556-562 [18] Ding C,Li T,Peng W,et al．Orthogonal nonnegative matrix tri-factorizations for clustering [C]∥Proceedings of 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining．2006:126-135 [19] Georghiades A S,Belhumeur P N,Kriegman D J,et al．From few to many:illumination cone models for face recognition under variable lighting and pose [J]．IEEE Trans．Pattern Anal．Mach．Intelligence,2001,3(6):643-660 [20] Sim T,Baker S,Bsat M．The CMU pose,illumination,and exprssion database[J]．IEEE Trans．Pattern Anal．Mach．Intell.,2003,5(12):1615-1618 |
| No related articles found! |
|
||
首页