For sparse coding, the weaker the correlation of dictionary atoms is, the better the representation
capacity of dictionary will be. A weak correlation dictionary construction method for sparse coding has been
proposed in this paper. Firstly, a new dictionary atom initialization is proposed in which data samples with weak
correlation are selected as the initial dictionary atoms in order to effectively reduce the correlation among them.
Then, in the process of dictionary learning, the correlation between atoms has been measured by correlation
coefficient, and strong correlation atoms have been eliminated and replaced by weak correlation atoms in order
to improve the representation capacity of the dictionary. An image classification scheme has been achieved by
applying the weak correlation dictionary construction method proposed in this paper. Experimental results show
that, the proposed method averagely improves image classification accuracy by more than 2%, compared to sparse
coding spatial pyramid matching (ScSPM) and other existing methods for image classification on the datasets of
Caltech-101, Scene-15, etc.
LONG Haixia (龙海霞), ZHUO Li* (卓 力), QU Panling (屈盼玲), ZHANG Jing (张 菁)
. Weak Correlation Dictionary Construction Method for Sparse Coding[J]. Journal of Shanghai Jiaotong University(Science), 2017
, 22(1)
: 77
-081
.
DOI: 10.1007/s12204-017-1803-y
[1] LEE H, BATTLE A, RAINA R, et al. Efficient sparsecoding algorithms [C]//Advances in Neural InformationProcessing Systems. Massachusetts: MIT press,2006: 801-808.
[2] BAO C L, QUAN Y H, JI H. A convergent incoherentdictionary learning algorithm for sparse coding[C]//Proceedings of Computer Vision-ECCV 2014.Berlin: Springer-Verlag, 2014: 302-316.
[3] MAILH′E B, BARCHIESI D, PLUMBLEY M D. INKSVD:Learning incoherent dictionaries for sparse representations[C]//IEEE International Conference onAcoustics, Speech and Signal Processing (ICASSP).[s.l.]: IEEE, 2012: 3573-3576.
[4] LIAN Q S, SHI B S, CHEN S Z. Research advanceon dictionary learning models, algorithms and applications[J]. Acta Automatica Sinica, 2015, 41(2): 240-260 (in Chinese).
[5] BARCHIESI D, PLUMBLEY M D. Learning incoherentdictionaries for sparse approximation using iterativeprojections and rotations [J]. IEEE Transactionson Signal Processing, 2013, 61(8): 2055-2065.
[6] MUKHERJEE L, HALL A. Non-negative sparse codingwith regularizer for image classification [C]//IEEEWinter Conference on Applications of Computer Vision(WACV). Waikoloa, USA: IEEE, 2015: 852-859.
[7] YANG J C, YU K, GONG Y H, et al. Linear spatialpyramid matching using sparse coding for image classification[C]//IEEE Conference on Computer Visionand Pattern Recognition. Miami, USA: IEEE, 2009:1794-1801.
[8] LAZEBNIK S, SCHMID C, PONCE J. Beyond bagsof features: Spatial pyramid matching for recognizingnatural scene categories [C]//IEEE Computer SocietyConference on Computer Vision and Pattern Recognition.Washington, USA: IEEE, 2006, 2: 2169-2178.
[9] XIE L X, TIAN Q, ZHANG B. Spatial poolingof heterogeneous features for image applications[C]//Proceedings of the 20th ACM International Conferenceon Multimedia. Nara, Japan: ACM, 2012: 539-548.
[10] LI F F, FERGUS R, PERONA P. Learning generativevisual models from few training examples: An incrementalBayesian approach tested on 101 object categories[J]. Computer Vision and Image Understanding,2007, 106(1): 59-70.
[11] LI F F, PERONA P. A Bayesian hierarchical model forlearning natural scene categories [C]//IEEE ComputerSociety Conference on Computer Vision and PatternRecognition. San Diego, USA: IEEE, 2005: 524-531.
[12] VAN GEMERT J C, GEUSEBROEK J M, VEENMANC J, et al. Kernel codebooks for scene categorization[C]//Proceedings of Computer Vision-ECCV.Berlin: Springer-Verlag, 2008: 696-709.
