Professional Documents
Culture Documents
Neurocomputing: Bao-Di Liu, Bin Shen, Liangke Gui, Yu-Xiong Wang, Xue Li, Fei Yan, Yan-Jiang Wang
Neurocomputing: Bao-Di Liu, Bin Shen, Liangke Gui, Yu-Xiong Wang, Xue Li, Fei Yan, Yan-Jiang Wang
Neurocomputing
journal homepage: www.elsevier.com/locate/neucom
art ic l e i nf o a b s t r a c t
Article history: Recently, sparse representation based classification (SRC) and collaborative representation based clas-
Received 2 March 2015 sification (CRC) have been successfully used for visual recognition and have demonstrated impressive
Received in revised form performance. Given a test sample, SRC or CRC formulates its linear representation with respect to the
7 July 2015
training samples and then computes the residual error for each class. SRC or CRC assumes that the
Accepted 20 August 2015
Available online 8 April 2016
training samples from each class contribute equally to the dictionary in the corresponding class, i.e., the
dictionary consists of the training samples in that class. This, however, leads to high residual error and
Keywords: instability. To overcome this limitation, we propose a class specific dictionary learning algorithm. To be
Class specific dictionary learning specific, by introducing the dual form of dictionary learning, an explicit relationship between the basis
Sparse representation
vectors and the original image features is represented, which also enhances the interpretability. SRC or
Collaborative representation
CRC can be thus considered as a special case of the proposed algorithm. Blockwise coordinate descent
Face recognition
algorithm and Lagrange multipliers are then adopted to optimize the corresponding objective function.
Extensive experimental results on five benchmark face recognition datasets show that the proposed
algorithm achieves superior performance compared with conventional classification algorithms.
& 2016 Elsevier B.V. All rights reserved.
1. Introduction samples. Ho et al. [9] and Tao et al. [23] proposed a nearest sub-
space method to assign the label of a test image by comparing its
Face recognition is a classical yet challenging research topic in reconstruction error for each category.
computer vision and pattern recognition [33]. Two stages are Under the nearest subspace [44, 45] framework, Wright et al.
usually considered for effective face recognition: (1) feature [27] proposed a sparse representation based classification (SRC)
extraction, (2) classifier construction and label prediction. For the system and achieved impressive performance. Given a test sample,
first stage, Turk and Pentland [25] proposed eigenfaces by per- sparse representation techniques represent it as a sparse linear
forming principal component analysis (PCA). He et al. [8] proposed combination of the training samples. The predicted label is
laplacianfaces to preserve local information. Belhumeur et al. [2] determined by the residual error from each class. Different from
suggested fisherfaces to maximize the ratio of between-class traditional decomposition frameworks like PCA, non-negative
scatter to within-class scatter. Yan et al. [28] proposed a multi- matrix factorization [39], and low-rank factorization [40], SRC
subregion based correlation filter bank algorithm to extract both allows coding under over-complete bases, and thus makes the
the global-based and local-based face features. For the latter stage, attained sparse codes capable of representing the data more
Richard et al. [19] proposed a nearest neighbor method to predict adaptively and flexibly. To analyze SRC, Zhang et al. [31] proposed
the label of a test image using its nearest neighbors in the training collaborative representation based classification (CRC) as an
alternative approach. CRC represents a test sample as the linear
combination of almost all the training samples. They found that it
n
Corresponding author. Tel.: þ 86 15764217948. is the collaborative representation rather than the sparse repre-
E-mail addresses: thu.liubaodi@gmail.com (B.-D. Liu),
stanshenbin@gmail.com (B. Shen), guiliangke@gmail.com (L. Gui),
sentation that makes the nearest subspace method powerful for
yuxiongw@cs.cmu.edu (Y.-X. Wang), lixue421@gmail.com (X. Li), classification. SRC, CRC, and their variants have been also used in
232842281@qq.com (F. Yan), yjwang@upc.edu.cn (Y.-J. Wang). other visual data sensing and analysis tasks, such as image
http://dx.doi.org/10.1016/j.neucom.2015.08.128
0925-2312/& 2016 Elsevier B.V. All rights reserved.
B.-D. Liu et al. / Neurocomputing 204 (2016) 198–210 199
classification [12], image inpainting [41], object detection [42], demonstrated performance of hybrid dictionary learning, it is still
image annotation [1], and transfer learning [43]. a challenge to balance between the shared dictionary and the class
Despite their promise, both SRC and CRC algorithms directly use specific dictionary.
the training samples as the dictionary for each class. By contrast, a In this paper, motivated by the superior performance of the SRC
well learned dictionary, especially by enforcing some discriminative and CRC algorithms and the class specific dictionary learning
criteria, can reduce the residual error greatly and achieve superior method, we propose class specific dictionary learning (CSDL) for
performance for classification tasks. Existing discriminative dic- both sparse representation based classifier (CSDL-SRC) and colla-
tionary learning approaches are mainly categorized into three borative representation based classifier (CSDL-CRC). Fig. 1 shows
types: shared dictionary learning, class specific dictionary learning, the framework of our proposed CSDL. The major distinction
and hybrid dictionary learning. In shared dictionary learning, each between our approach and the existing class specific dictionary
basis is associated to all the training samples. Mairal et al. [16] learning methods is that the existing methods directly optimize
proposed to learn a discriminative dictionary with a linear classifier the dictionary basis vectors (the “primal” form), whereas we
of coding coefficients. Liu et al. [15] learned a Fisher discriminative leverage a “dual” reformulation of dictionary learning and opti-
dictionary. Liu et al. [35, 36, 37] and Yu et al. [38] presented a graph mize the weights instead. Compared with the “primal” form
embedded dictionary learning method. Zhang and Li [32] proposed widely used by the existing methods, our novel “dual” form offers
a joint dictionary learning algorithm for face recognition. In class several benefits:
specific dictionary learning, each basis only corresponds to a single
class so that the class specific reconstruction error could be used for It provides an explicit relationship between the basis vectors
classification. Yang et al. [30] learned a dictionary for each class with and the original image features, thus enhancing the interpret-
sparse coefficients and applied it for face recognition. Sprechmann ability of the learned dictionary.
and Sapiro [22] also learned a dictionary for each class with sparse It is easy to be kernelized due to the separation of original data,
representation and used it in signal clustering. Castrodad and Sapiro which is difficult for the existing methods. The generalization to
[4] learned a set of action specific dictionaries with non-negative kernel spaces will be elaborated in Section 5.7.3.
penalty on both dictionary atoms and representation coefficients. Most of the existing class specific dictionary learning methods
Wang et al. [26] introduced mutual incoherence information to focus on introducing additional regularization terms, which
promote class specific dictionary learning in action recognition. could be easily incorporated into our dual formulation of class
Yang et al. [29] embedded the Fisher discriminative information specific dictionary learning to further improve the performance.
into class specific dictionary learning.
The shared dictionary learning approaches usually lead to a Our main contributions are threefold:
dictionary of small size and the discriminative information (i.e.,
the label information corresponding to coding coefficients) is We propose a novel class specific dictionary learning scheme
embedded into the dictionary learning framework. The class spe- that considers the weight of each sample when generating the
cific dictionary learning approaches usually focus on the classifier dictionary (i.e., subspace). The traditional CRC and SRC methods
construction aspect since each basis vector is fixed to a single class perform face recognition without training procedures (i.e., the
label. The combination of shared basis vectors and class specific training samples are directly used for predicting the labels). By
basis vectors is then learned in hybrid dictionary learning. Zhou contrast, our proposed method compensates this deficiency by
et al. [34] learned a hybrid dictionary with Fisher regularization on introducing class specific dictionary learning. It is applicable to
the coding coefficient. Gao et al. [6] learned a shared dictionary to both CRC and SRC. Furthermore, it is reasonable to assume
encode common visual patterns and learned a class specific dic- additionally that different samples contribute unevenly in
tionary to encode subtle visual differences among different cate- constructing the corresponding subspace. CRC or SRC can be
gories for fine-grained image representation. Liu et al. [12] pro- thus viewed as special cases of our proposed algorithm.
posed a hierarchical dictionary learning method to produce a We propose the dual form of dictionary learning to enhance the
shared dictionary and a cluster specific dictionary. In spite of the interpretability.
Fig. 1. The framework of our proposed class specific dictionary learning algorithm.
200 B.-D. Liu et al. / Neurocomputing 204 (2016) 198–210
class:
c 2
idðyÞ ¼ arg minc y X c s^ 2 : ð2Þ ⎡W 1 0 0 L ⎤ 0
c
s^ represents the partial sparse codes related to images in class c. ⎢ 2 ⎥
The procedure of SRC is shown in Algorithm 1. The impressive ⎢0 W 0 L 0 ⎥
results of SRC have been reported in Wright et al. [27]. ⎢0 0 W3 L 0 ⎥
Algorithm 1. Algorithm for SRC.
⎢ ⎥
⎢L L L L L ⎥
Require: Training samples X A RDN , α, and test sample y ⎢ ⎥
Code y with the dictionary X via ℓ1-minimization. ⎢⎣0 0 0 L WC⎥
1:
n 2 o ⎦
s^ ¼ arg mins y Xs þ 2αksk1
2
Fig. 2. The learned weight coefficient matrix W for constructing dictionary.
B.-D. Liu et al. / Neurocomputing 204 (2016) 198–210 201
Similarly, the objective function of learning dictionary for col- Based on the convexity and monotonic property of the parabolic
laborative representation is function, it is easy to know that f ðSckn Þ reaches the minimum at the
unique point
f W c ; Sc ¼ X c X c W c Sc 2F þ βSc 2F
c c 2
Sckn ¼ minf½W c X c X c kn ½ES~c kn ; αg
kn
s:t: X W •k F r 1; ∀k ¼ 1; 2; …; K: ð6Þ T T
one draws an analogy with linear SVM training: one can formulate
þ maxf½W c X c X c k ½ES~c k ; αg;
T T k
the training problem as an optimization over (1) either directly the ð11Þ
( c
weights of a linear classifier vector of the same dimension as the Sp ; p a k
where S~c ¼
k
input signal (2) or the support vector weights, that is a vector of
0; p ¼ k:
dimension equal to the size of the training set that is used to
linearly combine the training inputs. In the context of the problem 4.1.2. ℓ2–ℓs minimization subproblem
here, the support vector weights are analogous to the weights W. With Sc fixed , the objective function of ℓ2–ℓs subproblem
The atoms now capture a notion of centroids similar to K-means, becomes
which explicitly expresses what happens during dictionary
f W c ¼ X c X c W c Sc 2F s:t: X c W c•k 2F r 1; ∀k ¼ 1; 2; …; K: ð12Þ
learning, leading to enhanced interpretability.
Here, the Lagrange multipliers are used to solve the ℓ2-norm
constrained minimization subproblem.
4. Optimization of the objective function W c can be optimized in a column-wise manner. Specifically,
T
ignoring the constant term trfX c X c g, its Lagrangian is
In this section, we focus on solving the optimization problem of K h
X i
LðW c ; λk ; μk Þ ¼ 2
T
the two types of CSDL algorithms proposed above. Specifically, Sc X c X c W c•k
k
similar to the optimization strategy used in Lee et al. [11] and Liu k¼1
ð16Þ
n o N h
X i (
T T
f Sc ¼ tr X c X c 2 X c X c W c Sc•n W•pc; p≠k
where F ¼ Sc Sc and W~ c ¼
T k
n•
n¼1
0; p ¼ k:
X
N h i X
K X
N
þ
T T T
Sc•k W c X c X c W c Sc•n þ 2α Sc j: ð8Þ Algorithm 2. CSDL algorithm for sparse representation.
kn
n¼1 k¼1n¼1
Require Data matrix X c A RDN , α, and K
T
Ignoring the constant term trfX c X c g, the objective function of Sckn 1: W c ’randðN c ; KÞ; Sc ’zerosðK; N c Þ
c
reduces
to Eqn. (9) with W and S1n ; S2n ; …; SKn ⧹Sckn fixed. Here,
c c c
2: for k ¼ 1; k r K; k þ þ do
( Sc1n ; Sc2n ; …; ScKn ⧹Sckn stands for all elements in Sc except the qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3: T
element in the kth row and nth column): W c•k ¼ W c•k = W c•k W c•k
( ) 4: end for
nh i o XK h i
T T T T
f Sckn ¼ Sckn 2 W c X c X c W c kk þ2Sckn W c X c X c W c kl Scln 5: iter ¼ 0
l ¼ 1;l≠k 6: while ðf ðiterÞ f ðiter þ 1ÞÞ=f ðiterÞ 4 1e 5 do
nh i o 7: iter’iter þ 1
2Sckn cT
W X X cT c
þ 2αSckn j: ð9Þ
kn 8: Update Sc :
9: T T
T T
Here, ½W c X c X c W c kk ¼ 1. f ðSckn Þ is a piece-wise parabolic function. E ¼ W c Xc XcW c
202 B.-D. Liu et al. / Neurocomputing 204 (2016) 198–210
10: for k ¼ 1; k rK; k þ þ do For ℓ2–ℓs regularized minimization subproblem, the objective
11: function is
¼ minf½W X X k ½ES~c k ; αg
cT k
cT
Sck c
2 2
12:
þ maxf½W c X c X c k ½ES~c k ; αg
T T k
f ðSc Þ ¼ X c X c W c Sc F þ βSc 2 : ð18Þ
13: end for Eqn. (18) can be easily solved by derivation and its analytical
14: Update W c : solutions is
15: T
Compute F ¼ Sc Sc , G ¼ F ð1 IÞ 1
16: for k ¼ 1; k rK; k þ þ do Sc ¼ W Tc X Tc X c W c þ βI W Tc X Tc X c : ð19Þ
17: T
W c•k ¼ Sc•k W c G•k
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi For ℓ2–ℓs constrained minimization subproblem, the objective
18: T T
W c•k ¼ W c•k = W c•k X c X c W c•k function is the same as Eqn. (12), and the solution thus remains as
19: end for Eqn. (16).
20: Update the objective function:
21: P
Nc 4.3. CSDL for label prediction
f ¼ ‖X c X c W c Sc ‖2F þ 2α ‖Sc•n k1
n¼1
22: end while Algorithm 3 shows our proposed CSDL-SRC or CSDL-CRC algo-
23: return W c , and Sc rithm for predicting the label of a test sample.
Fig. 4. Influence of α for SRC & CSDL-SRC and β for CRC & CSDL-CRC on the Yale dataset, respectively.
Fig. 6. Influence of α for SRC & CSDL-SRC and β for CRC & CSDL-CRC on the ORL dataset, respectively.
Fig. 5 shows some sample images from the dataset. The per- ℓ2 norm Power norm
formance with different values of α or β is reported in Fig. 6.
Table 2 shows the recognition rate of NN, SVM, CRC, SRC, CSDL- NN 85.8 73.18 84.25 72.82
SVM 94.65 72.57 93.7 72.95
CRC, and CSDL-SRC. From Table 2, our proposed CSDL algorithm CRC 94.6 72.39 93.75 72.29
achieves superior performance to the other four classical clas- SRC 93.65 72.22 92.157 1.55
sification methods. CSDL-CRC 96.457 1.61 95.45 71.96
CSDL-SRC 95.55 71.50 95.507 1.45
For the Extended YaleB dataset, there are 2,414 frontal face 5.5. CMU PIE dataset
images and 38 individuals in total. All the images are captured
under varying illumination conditions. Fig. 7 shows some sample The CMU PIE dataset contains 41,368 images of 68 indivi-
images from the dataset. The performance with different values of duals in total. Each individual is under 13 different poses, 43
α or β is reported in Fig. 8. Table 3 shows the recognition rate of different illumination conditions, and with 4 different expres-
NN, SVM, CRC, SRC, CSDL-CRC, and CSDL-SRC. From Table 3, our sions. Each individual thus may lie on multiple manifolds. Five
proposed CSDL algorithm achieves superior performance to the near frontal poses (C05, C07, C09, C27, C29) and all different
other four classical classification methods. illuminations and expressions are used in our experiment. There
B.-D. Liu et al. / Neurocomputing 204 (2016) 198–210 205
Fig. 8. Influence of α for SRC & CSDL-SRC and β for CRC & CSDL-CRC on the Extended YaleB dataset, respectively.
are about 170 images for each individual and 11,554 images in
total. Fig. 9 shows some sample images from the dataset. The Table 3
performance with different values of α or β is reported in Fig. 10. Recognition rate on the Extended YaleB dataset (%).
Table 4 shows the recognition rate of NN, SVM, CRC, SRC, CSDL-
Methods Features
CRC, and CSDL-SRC. From Table 4, our proposed CSDL algorithm
achieves superior performance to the other four classical clas- ℓ2 norm Power norm
sification methods.
NN 53.147 1.24 51.46 71.49
SVM 85.767 0.80 94.82 70.76
5.6. AR dataset CRC 89.96 7 1.04 97.377 0.74
SRC 89.92 7 1.09 97.13 70.56
For the AR dataset, there are over 4,000 frontal faces for 126 CSDL-CRC 90.36 7 0.90 97.747 0.70
individuals. A subset consisting of 50 male and 50 female cate- CSDL-SRC 91.467 0.64 98.287 0.57
gories is used here. There are 26 face images for each class.
Compared with the two former datasets, the AR dataset contains
more facial variations, such as illumination change, various comparison with state-of-the-art dictionary learning based face
expressions, and facial disguises. Fig. 11 shows some samples face recognition methods is presented.
images from the dataset. The performance with different values of
α or β is reported in Fig. 12. Table 5 shows the recognition rate of 5.7.1. The size of bases K per class
NN, SVM, CRC, SRC, CSDL-CRC, and CSDL-SRC. From Table 5, our Fig. 13 shows the influence of the size of bases K per class for
proposed CSDL algorithm achieves superior performance to the our proposed CSDL-CRC and CSDL-SRC. From Fig. 13, when K is
other four classical classification methods. small, e.g., K ¼5, the performance of CSDL is relative poor. When K
is larger than twice of the number of the training samples per
5.7. Analysis and discussions class, the performance becomes high and steady.
In this section, first, the size of bases per class is briefly ana- 5.7.2. Visualization of the learned dictionary
lyzed. Second, the learned dictionary is visualized and analyzed. Fig. 14 shows the comparison of the original face images and
Third, the generalization to kernel space is illustrated. Finally, the class specific learned dictionary with the size 20 (the learned
206 B.-D. Liu et al. / Neurocomputing 204 (2016) 198–210
Fig. 10. Influence of α for SRC & CSDL-SRC and β for CRC & CSDL-CRC on the CMU PIE dataset.
Fig. 12. Influence of α for SRC & CSDL-SRC and β for CRC & CSDL-CRC on the AR dataset, respectively.
n o
ℓ2 norm Power norm T
2 tr κ X c ; X c W c Sc þ tr Sc W c κ X c ; X c W c Sc þ βSc 2H :
T
Fig. 14. Comparison of the original faces and the learned dictionary from the Extended YaleB dataset with ℓ2 normalized features.
Acknowledgment [26] H. Wang, C. Yuan, W. Hu, C. Sun, Supervised class-specific dictionary learning
for sparse modeling in action recognition, Pattern Recognit. 45 (11) (2012)
3902–3911.
This paper is supported partly by the National Natural Science [27] J. Wright, A.Y. Yang, A. Ganesh, S. Sastry, Y. Ma, Robust face recognition via
Foundation of China (Grant nos. 61402535, 61271407), the Natural sparse representation, IEEE Trans. Pattern Anal. Mach. Intell. 31 (2) (2009)
210–227.
Science Foundation for Youths of Shandong Province, China (Grant
[28] Y. Yan, H. Wang, D. Suter, Multi-subregion based correlation filter bank for
no. ZR2014FQ001) , Qingdao Science and Technology Project (no. robust face recognition, Pattern Recognit. 47 (11) (2014) 3487–3501.
14-2-4-111-jch), and the Fundamental Research Funds for the [29] M. Yang, L. Zhang, X. Feng, D. Zhang, Sparse representation based fisher dis-
crimination dictionary learning for image classification, Int. J. Comput. Vis. 109
Central Universities, China University of Petroleum (East China) (3) (2014) 209–232.
(Grant no. 14CX02169A). [30] M. Yang, L. Zhang, J. Yang D. Zhang, Metaface learning for sparse representa-
tion based face recognition, in: Proceedings of the 17th International Con-
ference on Image Processing, IEEE, Hong Kong, 2010, pp. 1601–1604.
[31] D. Zhang, M. Yang, X. Feng, Sparse representation or collaborative repre-
sentation: which helps face recognition? in: Proceedings of the 13th Inter-
References national Conference on Computer Vision, IEEE, Barcelona, 2011, pp. 471–478.
[32] Q. Zhang, B. Li, Discriminative k-svd for dictionary learning in face recognition,
[1] D. Tao, L. Jin, W. Liu, X. Li, Hessian regularized support vector machines for in: Proceedings of the 23rd International Conference on Computer Vision and
mobile image annotation on the cloud, IEEE Trans. on Multimedia 15 (4) Pattern Recognition, IEEE, San Francisco, California, 2010, pp. 2691–2698.
(2013) 833–844. [33] W. Zhao, R. Chellappa, P.J. Phillips, A. Rosenfeld, Face recognition: a literature
[2] P.N. Belhumeur, J.P. Hespanha, D. Kriegman, Eigenfaces vs. fisherfaces: survey, ACM Comput. Surv. 35 (4) (2003) 399–458.
recognition using class specific linear projection, IEEE Trans. Pattern Anal. [34] N. Zhou, Y. Shen, J. Peng, J. Fan, Learning inter-related visual dictionary for
Mach. Intell. 19 (7) (1997) 711–720. object recognition, in: Proceedings of the 25th International Conference on
[3] D.P. Bertsekas, Nonlinear Programming, Athena Scientific, Belmont, MA, 1999. Computer Vision and Pattern Recognition, IEEE, Province, Rhode Island, 2012,
[4] A. Castrodad, G. Sapiro, Sparse modeling of human actions from motion pp. 3490–3497.
imagery, Int. J. Comput. Vis. 100 (1) (2012) 1–15. [35] W. Liu, D. Tao, Multiview hessian regularization for image annotation, IEEE
[5] C. Chang, C. Lin, Libsvm: a library for support vector machines, ACM Trans. Trans. Image Process 22 (7) (2013) 2676–2687.
Intell. Syst. Technol. 2 (3) (2011) 27. [36] W. Liu, D. Tao, J. Cheng, Y. Tang, Multiview hessian discriminative sparse coding
[6] S. Gao, I.W.-H. Tsang, Y. Ma, Learning category-specific dictionary and shared for image annotation, Comput. Vis. Image Underst 118 (1) (2014) 50–60.
dictionary for fine-grained image categorization, IEEE Trans. Image Process. 23 [37] W. Liu, Z. Zha, Y. Wang, K. Lu, D. Tao, p-Laplacian regularized sparse coding for
(2) (2014) 623–634. human activityr recognition, IEEE Trans. Ind. Electron. (2016).
[7] A. Georghiades, P. Belhumeur, D. Kriegman, From few to many: illumination [38] J. Yu, D. Tao, M. Wang, Adaptive hypergraph learning and its application in
cone models for face recognition under variable lighting and pose, IEEE Trans. image classification, IEEE Trans. Image Process 21 (7) (2012) 3262–3272.
Pattern Anal. Mach. Intell. 23 (6) (2001) 643–660. [39] Y.-X. Wang, Y.-J. Zhang, Nonnegative matrix factorization: a comprehensive
[8] X. He, S. Yan, Y. Hu, P. Niyogi, H.-J. Zhang, Face recognition using laplacianfaces, review, IEEE Trans. Know. Data Eng 25 (6) (2013) 1336–1353.
IEEE Trans. Pattern Anal. Mach. Intell. 27 (3) (2005) 328–340. [40] Y.-X. Wang, L.-Y. Gui, Y.-J. Zhang, Neighborhood preserving non-negative
[9] J. Ho, M. Yang, J. Lim, K. Lee, D. Kriegman, Clustering appearances of objects tensor factorization for image representation, in: Proceedings of the 37th IEEE
under varying illumination conditions, in: Proceedings of 16th International International Conference on Acoustics, Speech and Signal Processing, Kyoto,
Conference on Computer Vision and Pattern Recognition, IEEE, Madison, 2012, pp. 3389–3392.
Wisconsin, 2003, pp. 1–11. [41] Y.-X. Wang, Y.-J. Zhang, Image inpainting via weighted sparse non-negative
[10] Z. Jiang, Z. Lin, L.S. Davis, Label consistent k-svd: learning a discriminative matrix factorization, in: Proceedings of the 18th IEEE International Conference
dictionary for recognition, IEEE Trans. Pattern Anal. Mach. Intell. 35 (11) (2013) on Image Processing, Brussels, 2011, pp. 3409–3412.
2651–2664. [42] Y.-X. Wang, M. Hebert, Model recommendation: Generating object detectors
[11] H. Lee, A. Battle, R. Raina, A. Ng, Efficient sparse coding algorithms, in: Pro- from few samples, in: Proceedings of the 28th IEEE International Conference
ceedings of Advances in Neural Information Processing Systems, The MIT on Computer Vision and Pattern Recognition, Boston, Massachusetts, 2015, pp.
Press, Vancouver, British Columbia, 2006, pp. 801–808. 1619–1628.
[12] B.-D. Liu, B. Shen, X. Li, Locality sensitive dictionary learning for image clas- [43] Y.-X. Wang, M. Hebert, Learned by transferring from unsupervised universal
sification, in: Proceedings of the 22nd International Conference on Image sources, in: Proceedings of the 30th AAAI Conference on Artificial Intelligence,
Processing, IEEE, Quebec, 2015, pp. 3807-3811.. Phoenix, Arizona, 2016.
[13] B.-D. Liu, B. Shen, Y.-X. Wang, Class specific dictionary learning for face [44] C. Xu, D. Tao, C. Xu, Multi-view intact space learning, IEEE Trans. Pattern Anal.
recognition, in: Proceedings of the International Conference on Security, Pat- Mach. Intell 37 (12) (2015) 2531–2544.
tern Analysis, and Cybernetics, IEEE, Wuhan, Hubei, 2014, pp. 229–234. [45] C. Xu, D. Tao, C. Xu, Large-Margin Multi-view information bottleneck, IEEE
[14] B.-D. Liu, Y.-X. Wang, S. Bin, Y.-J. Zhang, Y.-J. Wang, Blockwise coordinate Trans. Pattern Anal. Mach. Intell 36 (8) (2014) 1559–1572.
descent schemes for sparse representation, in: Proceedings of the 39th
International Conference on Acoustics, Speech and Signal Processing, IEEE,
Florence, 2014, pp. 5304–5308.
[15] B.-D. Liu, Y.-X. Wang, Y.-J. Zhang, Y. Zheng, Discriminant sparse coding for Bao-Di Liu received the Ph.D. degree in Electronic
image classification, in: Proceedings of the 37th International Conference on Engineering from Tsinghua University. Currently, he is
Acoustics, Speech and Signal Processing, IEEE, Kyoto, 2012, pp. 2193–2196. an assistant professor in College of Information and
[16] J. Mairal, J. Ponce, G. Sapiro, A. Zisserman, F.R. Bach, Supervised dictionary Control Engineering, China University of Petroleum,
learning, in: Proceedings of Advances in Neural Information Processing Sys- China. His research interests include computer vision
tems, The MIT Press, Vancouver, British Columbia, 2009, pp. 1033–1040. and machine learning.
[17] A. Martinez, The AR Face Database, CVC Technical Report 24, 1998.
[18] I. Ramirez, P. Sprechmann, G. Sapiro, Classification and clustering via dic-
tionary learning with structured incoherence and shared features, in: Pro-
ceedings of the 23rd International Conference on Computer Vision and Pattern
Recognition, IEEE, San Francisco, California, 2010, pp. 3501–3508.
[19] O.D. Richard, E.H. Peter, G.S. David, Pattern classification. A Wiley-Interscience,
2001, pp. 373–378.
[20] F.S. Samaria, A.C. Harter, Parameterisation of a stochastic model for human
face identification, in: Proceedings of the 2nd Workshop on Applications of
Computer Vision, IEEE, Sarasota, Florida, 1994, pp. 138-142.
[21] T. Sim, S. Baker, M. Bsat, The cmu pose, illumination, and expression (pie) Bin Shen is a Ph.D candidate in Department of Com-
database, in: Proceedings of the fifth International Conference on Automatic puter Science, Purdue University, West Lafayette, IN
Face and Gesture Recognition, IEEE, Shanghai, 2002, pp. 46–51. 47907, USA. Before joining Purdue, he got B.S. and M.S.
[22] P. Sprechmann, G. Sapiro, Dictionary learning and sparse coding for unsu- degrees from EE, Tsinghua University, Beijing, in 2007
pervised clustering, in: Proceedings of the 35th International Conference on and 2009, respectively. His research interests include
Acoustics, Speech and Signal Processing, IEEE, Dallas, Texas, 2010, pp. 2042– image processing, machine learning and data mining.
2045.
[23] D. Tao, X. Li, X. Wu, S.J. Maybank, Geometric mean for subspace selection, IEEE
Trans. Pattern Anal. Mach. Intell. 31 (2) (2009) 260–274.
[24] P. Tseng, Convergence of a block coordinate descent method for non-
differentiable minimization, J. Optim. Theory Appl. 109 (3) (2001) 475–494.
[25] M. Turk, A. Pentland, Eigenfaces for recognition, J. Cogn. Neurosci. 3 (1) (1991)
71–86.
210 B.-D. Liu et al. / Neurocomputing 204 (2016) 198–210
Yu-Xiong Wang is a Ph.D. student in the Robotics Fei Yan, received the B.S. degree in Electronic and
Institute, School of Computer Science, at Carnegie Information Engineering, China University of Petro-
Mellon University. His research interests include leum. Currently, he is a staff in Lijin County Party
computer vision, image processing, and machine Committee Office, China. His research interests include
learning. computer vision and machine learning.
Xue Li received the B.S. degree in Electronic Engineer- Yan-Jiang Wang received the M.S. degree from Beijing
ing from Beijing Institute of Technology (BIT), Beijing, University of Aeronautics and Astronautics, Beijing,
China, in 2011. Currently, she is a Ph.D candidate in the China, in 1989 and the Ph.D. degree from Beijing Jiao-
Department of Electronic Engineering at Tsinghua tong University, Beijing, China, in 2001. Now he is a
University, Beijing, China. Her research interests professor of the College of Information and Control
include image classification, automatic image annota- Engineering, China University of Petroleum, Qingdao,
tion and machine learning. China. He is also the head of the Institute of Signal and
Information Processing, China University of Petroleum.
His research interests include pattern recognition,
computer vision, and cognitive computation.