论文信息 - Sparse representation of images using alternating linear programming

Sparse representation of images using alternating linear programming

Based on nonnegative matrix factorization, a set of images are represented by a product of two nonnegative matrices, over-complete basis matrix (features) and nonnegative coefficient matrix (sparse coding) in this paper. Under the objective that both basis matrix and coefficient matrix are sparse, an alternating linear programming (ALP) algorithm is proposed. And the ALP algorithm is proved to be convergent. After the very large scale alternating linear programming problems are converted to equivalent sets of linear programming subproblems, they can be solved much more efficiently. Furthermore, the ALP algorithm is extended and generalized to an alternating iterative optimization (AIO) algorithm. At last, simulation results show the validity of the proposed approach.

Yuanqing Li | Andrzej Cichocki | A. Cichocki | Yuanqing Li

[1] H. Sebastian Seung,et al. Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[2] Bruno A. Olshausen,et al. Sparse Codes and Spikes , 2001 .

[3] Joseph F. Murray,et al. An Improved FOCUSS-Based Learning Algorithm for Solving Blind Sparse Linear Inverse Problems , 2001 .

[4] Terrence J. Sejnowski,et al. Learning Overcomplete Representations , 2000, Neural Computation.

[5] Joseph F. Murray,et al. An improved FOCUSS-based learning algorithm for solving sparse linear inverse problems , 2001, Conference Record of Thirty-Fifth Asilomar Conference on Signals, Systems and Computers (Cat.No.01CH37256).

[6] Michael A. Saunders,et al. Atomic Decomposition by Basis Pursuit , 1998, SIAM J. Sci. Comput..

[7] David J. Field,et al. Sparse coding with an overcomplete basis set: A strategy employed by V1? , 1997, Vision Research.

[8] Bruno A. Olshausen,et al. Learning Sparse Image Codes using a Wavelet Pyramid Architecture , 2000, NIPS.