Batch Dictionary Learning with Augmented Orthogonal Matching Pursuit
Dictionary learning is often incorporated in classification method, which can obtain a new representation under the learned dictionary to achieve better classification performance. In this paper, we propose a novel Batch Dictionary Learning model with augmented orthogonal matching pursuit classification. Batch Dictionary Learning model is capable of improving the dictionary by removing the redundancy of over-complete dictionary, thus the learned optimal dictionary is more suitable for classification. To solve the optimization target, we improve the traditional orthogonal matching pursuit (OMP) algorithm and propose an augmented orthogonal matching pursuit algorithm (AOMP) to solve the objective function. Superior experimental results demonstrate that our proposed model outperform the other state-of-the-art classification algorithms on real-world dataset.
KeywordsClassification Sparse learning Dictionary learning Sparse representation
This research was supported in part by the Chinese National Natural Science Foundation under Grant nos. 61402395, 61472343 and 61502412, Natural Science Foundation of Jiangsu Province under contracts BK20140492, BK20151314, Jiangsu overseas research and training program for university prominent young and middle-aged teachers and presidents, Jiangsu government scholarship funding.
- 2.Engan, K., Aase, S.O., Husoy, J.H.: Method of optimal directions for frame design. In: Proceedings of 1999 IEEE International Conference on Acoustics, Speech, and Signal Processing, vol. 5, pp. 2443–2446. IEEE (1999)Google Scholar
- 4.Aharon, M., Elad, M., Bruckstein, A.: K-SVD: design of dictionaries for sparse representation. In: Proceedings of SPARS, vol. 5, pp. 9–12 (2005)Google Scholar
- 6.Farouk, R.M., Elsayed, M., Aly, M.: Medical image denoising based on log-Gabor wavelet dictionary and K-SVD algorithm. Int. J. Comput. Appl. 141(2016)Google Scholar
- 8.Chen, J., Liu, S., Huang, M., et al.: Dictionary learning for MRI denoising based on modified K-SVD. J. Imaging Sci. Technol. 61: 30505-1–30505-10 (2017)Google Scholar