改进的鉴别稀疏保持投影人脸识别算法
|
摘要
为了解决判别稀疏邻域保持嵌入(DSNPE)算法中时间复杂度偏高的问题,构造了一种新类间离散度.用各类样本的平均向量组成过完备字典去重构表达每一类的平均向量,然后通过最大间距准则(MMC)构造新的目标函数,更好地展现人脸样本数据库类间的差异,增强了类间判别力和鲁棒性,简化了字典和字典表达,降低算法复杂度.实验结果表明:改进后的算法在保持识别率优势的前提下,极大地减少了识别时间.
The improved formulation of between-neighborhood scatter was proposed in order to try to solve the problem on time in the algorithm discriminant sparse locality and preserving projections(DSNPE). It comes to select the mean vector set of each class as the over complete dictionary to represent the mean vector of each class. And then we formed a new target function through maximum margin criterion(MMC). As a result,the improved algorithm reveals the between-class distance exactly,improves the robustness,simplifies the dictionary and representation,decreases complexity of the algorithm.The experiment manifests that the improved algorithm shortens the proceeding time heavily under the guarantee of recognition rate.
The improved formulation of between-neighborhood scatter was proposed in order to try to solve the problem on time in the algorithm discriminant sparse locality and preserving projections(DSNPE). It comes to select the mean vector set of each class as the over complete dictionary to represent the mean vector of each class. And then we formed a new target function through maximum margin criterion(MMC). As a result,the improved algorithm reveals the between-class distance exactly,improves the robustness,simplifies the dictionary and representation,decreases complexity of the algorithm.The experiment manifests that the improved algorithm shortens the proceeding time heavily under the guarantee of recognition rate.
引文
[1]Fukunaga K.Introduction to statistical pattern recognition[M].2nd Edition.Boston:Boston Academic Press,1990.
[2]Duda R O,Hart P E,Stork D G.Pattern classification,seconded[M].New York:John Wiley&Sons,2000.
[3]Sch?lkopf B,Smola A,Müller K R.Nonlinear component analysis as a kernel eigenvalue problem[J].Neural Compution,1998,10(5):1299-1319.
[4]Tenenbaum J B,Silva V D,Langford J C.A global geometric framework for nonlinear dimensionality reduction[J].Science,2000,290:2319-2323.
[5]Belkin M,Niyogi P.Laplacian eigenmaps for dimensionality reduction and data representation[J].Neural Computation,2003,15(6):1373-1396.
[6]He X,Cai D,Yan S,et al.Neighborhood preserving embedding[C]//Proceedings in International Conference on Computer Vision(ICCV).[s.n.],2005:1208-1213.
[7]He X,Yan S,Hu Y,et al.Face recognition using Laplacian faces[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(3):328-340.
[8]Yu Weiwei.Two-dimensional discriminant locality preserving projections for face recognition[J].Image and Vision Computing,2009,30(15):1378-1383
[9]Wright J,Yang A,Sastry S,et al.Robust face recognition via sparse representation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(2):210-227.
[10]Qiao L,Chen Songcan,Tan X.Sparsity preserving projections with applications to face recognition[J].Pattern Recognition,2010,43(1):331-341.
[11]Lu Gui Fu,Jin Zhong,Zou Jian.Face recognition using discriminant sparsity neighborhood preserving embedding[J].Knowledge-Based Systems,2012,31(7):119-127.
[12]Li H,Jiang T,Zhang K.Efficient and robust feature extraction by maximum margin criterion[J].IEEE Transactions on Neural Networks,2006,17(1):1157-1165.
[13]Yang A Y,Sastry S S.Fast l1-minimization algorithms and an application in robust face recognition:a review[C]//Proceedings of 2010 IEEE 17th International Conference on Image Processing.[s.n.],2010:1849-1852.
[14]Donoho D L,Tsaig Y.Fast solution of L1-norm minimization problems when the solution may be sparse[J].IEEE Transactions on Information Theory,2008,54(11):4789-4812.
[15]Jia Yangqing,Nie Feiping,Zhang Changshui.Trace ratio problem revisited[J].IEEE Transactions on Neural Networks,2009,20(4):729-735.
[2]Duda R O,Hart P E,Stork D G.Pattern classification,seconded[M].New York:John Wiley&Sons,2000.
[3]Sch?lkopf B,Smola A,Müller K R.Nonlinear component analysis as a kernel eigenvalue problem[J].Neural Compution,1998,10(5):1299-1319.
[4]Tenenbaum J B,Silva V D,Langford J C.A global geometric framework for nonlinear dimensionality reduction[J].Science,2000,290:2319-2323.
[5]Belkin M,Niyogi P.Laplacian eigenmaps for dimensionality reduction and data representation[J].Neural Computation,2003,15(6):1373-1396.
[6]He X,Cai D,Yan S,et al.Neighborhood preserving embedding[C]//Proceedings in International Conference on Computer Vision(ICCV).[s.n.],2005:1208-1213.
[7]He X,Yan S,Hu Y,et al.Face recognition using Laplacian faces[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(3):328-340.
[8]Yu Weiwei.Two-dimensional discriminant locality preserving projections for face recognition[J].Image and Vision Computing,2009,30(15):1378-1383
[9]Wright J,Yang A,Sastry S,et al.Robust face recognition via sparse representation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(2):210-227.
[10]Qiao L,Chen Songcan,Tan X.Sparsity preserving projections with applications to face recognition[J].Pattern Recognition,2010,43(1):331-341.
[11]Lu Gui Fu,Jin Zhong,Zou Jian.Face recognition using discriminant sparsity neighborhood preserving embedding[J].Knowledge-Based Systems,2012,31(7):119-127.
[12]Li H,Jiang T,Zhang K.Efficient and robust feature extraction by maximum margin criterion[J].IEEE Transactions on Neural Networks,2006,17(1):1157-1165.
[13]Yang A Y,Sastry S S.Fast l1-minimization algorithms and an application in robust face recognition:a review[C]//Proceedings of 2010 IEEE 17th International Conference on Image Processing.[s.n.],2010:1849-1852.
[14]Donoho D L,Tsaig Y.Fast solution of L1-norm minimization problems when the solution may be sparse[J].IEEE Transactions on Information Theory,2008,54(11):4789-4812.
[15]Jia Yangqing,Nie Feiping,Zhang Changshui.Trace ratio problem revisited[J].IEEE Transactions on Neural Networks,2009,20(4):729-735.