A supervised multimanifold method with locality preserving for face recognition using single sample per person

Abstract

Although real-world experiences show that preparing one image per person is more convenient, most of the appearance-based face recognition methods degrade or fail to work if there is only a single sample per person (SSPP). In this work, we introduce a novel supervised learning method called supervised locality preserving multimanifold (SLPMM) for face recognition with SSPP. In SLPMM, two graphs: within-manifold graph and between-manifold graph are made to represent the information inside every manifold and the information among different manifolds, respectively. SLPMM simultaneously maximizes the between-manifold scatter and minimizes the within-manifold scatter which leads to discriminant space by adopting locality preserving projection (LPP) concept. Experimental results on two widely used face databases FERET and AR face database are presented to prove the efficacy of the proposed approach.

Keywords

face recognition / locality preserving / manifold learning / single sample per person

References

[1] TurkM, PentlandA. Eigenfaces for recognition [J]. Journal of Cognitive Neuroscience, 1991, 3: 71-86

[2] BelhumeurP N, HespanhaJ P, KriegmanD J. Eigenfaces vs. Fisherfaces: Recognition using class specific linear projection [J]. IEEE Transaction Pattern Analysis and Machine Intelligence, 1997, 19: 711-720

[3] TanX, ChenS, ZhauZ H, ZhangZ. Face recognition from a single image per person: A survey [J]. Pattern Recognition, 2006, 39: 1725-1745

[4] WuJ, ZhouZ. Face recognition with one training image per person [J]. Pattern Recognition Letters, 2002, 23: 1711-1719

[5] ChenS, ZhangD, ZhouZ. Enhanced (PC)2A for face recognition with one training image per person [J]. Pattern Recognition Letters, 2004, 25: 1173-1181

[6] YangJ, ZhangD, FrangiA, YangJ. Two-dimensional PCA: A new approach to appearance-based face representation and recognition [J]. IEEE Transaction Pattern Analysis and Machine Intelligence, 2004, 26: 131-137

[7] DengW, HuJ, GuoJ, CaiW, FengD. Robust, accurate and efficient face recognition from a single training image: A uniform pursuit approach [J]. Pattern Recognition, 2010, 43: 1748-1762

[8] WangJ, PlataniotisK N, VenetsanopoulosA N. On solving the face recognition problem with one training sample per subject [J]. Pattern Recognition, 2006, 39: 1746-1762

[9] CanM, ShanS, SuY, ChenX, GaoW. Adaptive discriminant analysis for face recognition from a single sample per person [C]. Automatic Face & Gesture Recognition and Workshops, 2011, Santa Barbara, CA, IEEE Press: 193199

[10] CanM, ShanS, SuY, XuD, ChenX. Adaptive discriminant learning for face recognition [J]. Pattern Recognition, 2013, 46: 2497-2509

[11] YinH, FuP, MengS H. Sampled FLDA for face recognition with single training image per person [J]. Pattern Recognition, 2006, 39: 2443-2445

[12] ZhangD, ChenS, ZhouZ. A new face recognition method based on SVD perturbation for single example image per person [J]. Applied Math and Computation, 2005, 163: 895-907

[13] GaoQ, ZhangL, ZhangD. Face recognition using FLDA with single training image per person [J]. Appl Math Comput, 2008, 205: 726-734

[14] ChenS, LiuJ, ZhouZ. Making FLDA applicable to face recognition with one sample per person [J]. Pattern Recognition, 2004, 37: 1553-1555

[15] TanX, ChenS, ZhouZ, ZhangF. Recognizing partially occluded, expression variant faces from single training image per person with SOM and soft K-NN ensemble [J]. IEEE Trans Neural Networks, 2005, 16: 875-886

[16] KananH, FaezK, GaoY. Face recognition using adaptively weighted patch PZM array from a single exemplar image per person [J]. Pattern Recognition, 2008, 41: 3799-3812

[17] TenenbaumJ B, SilvaV, LangfordJ. A global geometric framework for nonlinear dimensionality reduction [J]. Science, 2000, 290: 2319-2323

[18] RoweisS, SaulL. Nonlinear dimensionality reduction by locally linear embedding [J]. Science, 2000, 290: 2323-2326

[19] LuJ, TanY, VangG. Discriminative multimanifold analysis for face recognition from a single training sample per person [J]. IEEE Transaction on Pattern Analysis and Machine Intelligence, 2013, 35: 39-51

[20] NabipourM, AghagolzadehA, MotameniH. A DCT-based multimanifold face recognition method using single sample per person [C]. The 8th Symposium on Advances in Science and Technology, 2014

[21] NabipourM, AghagolzadehA, MotameniH. Multimanifold analysis with adaptive neighborhood in DCT domain for face recognition using single sample per person [C]. 22nd Iranian Conference on Electrical Engineering, 2014, Tehran, Iran, ICEE: 925930

[22] HeX, YanS, HuY, NiyogiP, ZhangH J. Face recognition using Laplacian faces [J]. IEEE Trans Pattern Analysis and Machine Intelligence, 2005, 27: 328-340

[23] YangW, SunC H, ZhangL. A multi-manifold discriminant analysis method for image feature extraction [J]. Pattern Recognition, 2011, 44: 1649-1657

[24] PhillipsP, WechslerH, HuangJ, RaussP. The FERET database and evaluation procedure for face recognition algorithms [J]. Image and Vision Computing, 1998, 16: 295-306

[25] MartinezA M, BenaventeRThe AR face database [R], 1998