Face recognition method based on HOG and DMMA from single training sample

Single training sample for face recognition technology has become one of the important topics in the field of computer vision, and presents various challenges and opportunities. In this paper, an improved method based on discriminative multi-manifold analysis (DMMA) algorithm is proposed for the single training sample problem in face recognition. The major contributions of the paper are that a novel method by fusing Histogram of the Oriented Gradient (HOG) features and DMMA algorithm is proposed, and a new adaptive method is applied to calculate similarity between patches of the face image. First, each face image is partitioned into several non-overlapping patches to form an image set for each sample per person. Then HOG operator is used to extract image histogram of each an image set, and the histogram of each an image set forms a statistics manifold. Finally, DMMA algorithm is applied to obtain the low-dimensional face image features, and the reconstruction-based manifold-manifold distance is used to identify the unlabeled faces. The performance of the proposed method is verified on the FERET and AR face databases. Experimental results indicate that the proposed method is superior to the general DMMA recognition algorithms under illumination variation and geometry transformation.

[1]  Shuicheng Yan,et al.  Graph Embedding and Extensions: A General Framework for Dimensionality Reduction , 2007 .

[2]  J. You,et al.  Enlarge the Training Set Based on Inter-Class Relationship for Face Recognition from One Image per Person , 2013, PloS one.

[3]  Jianxin Wu,et al.  Face recognition with one training image per person , 2002, Pattern Recognit. Lett..

[4]  M. Turk,et al.  Eigenfaces for Recognition , 1991, Journal of Cognitive Neuroscience.

[5]  David Zhang,et al.  Face recognition using FLDA with single training image per person , 2008, Appl. Math. Comput..

[6]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[7]  Ewa Kijak,et al.  Face recognition using Co-occurrence Histograms of Oriented Gradients , 2012, 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[8]  Chi Fang,et al.  Histogram of the oriented gradient for face recognition , 2011 .

[9]  A. Martínez,et al.  The AR face databasae , 1998 .

[10]  Guan-Chun Luh.  Face recognition using PCA based immune networks with single training sample per person , 2011, 2011 International Conference on Machine Learning and Cybernetics.

[11]  Changhui Hu,et al.  An adaptive approximation image reconstruction method for single sample problem in face recognition using FLDA , 2014, Multimedia Tools and Applications.

[12]  David J. Kriegman,et al.  Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection , 1996, ECCV.

[13]  J. Zhang,et al.  Global exponential stability of a class of neural networks with variable delays , 2005 .

[14]  S T Roweis,et al.  Nonlinear dimensionality reduction by locally linear embedding. , 2000, Science.

[15]  J. Tenenbaum,et al.  A global geometric framework for nonlinear dimensionality reduction. , 2000, Science.

[16]  Jiang-She Zhang,et al.  Face recognition using Elasticfaces , 2012, Pattern Recognit..

[17]  Zhi-Hua Zhou,et al.  Robust Face Recognition from a Single Training Image per Person with Kernel-Based SOM-Face , 2004, ISNN.

[18]  Yuxiao Hu,et al.  Face recognition using Laplacianfaces , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[19]  Karim Faez,et al.  Face recognition using adaptively weighted patch PZM array from a single exemplar image per person , 2008, Pattern Recognit..

[20]  Alejandro F. Frangi,et al.  Two-dimensional PCA: a new approach to appearance-based face representation and recognition , 2004 .

[21]  Shih-Fu Chang,et al.  Graph construction and b-matching for semi-supervised learning , 2009, ICML '09.

[22]  Mikhail Belkin,et al.  Laplacian Eigenmaps for Dimensionality Reduction and Data Representation , 2003, Neural Computation.

[23]  Daoqiang Zhang,et al.  Enhanced (PC)2 A for face recognition with one training image per person , 2004, Pattern Recognit. Lett..

[24]  Gang Wang,et al.  Discriminative multi-manifold analysis for face recognition from a single training sample per person , 2011, 2011 International Conference on Computer Vision.

[25]  Daoqiang Zhang,et al.  A new face recognition method based on SVD perturbation for single example image per person , 2005, Appl. Math. Comput..