Performance of the low rank matrix technique in image de-noising

 
 
 
  • Abstract
  • Keywords
  • References
  • Untitled
  • PDF
  • Abstract


    Generally, the technique of low rank matrix (LRM) estimation is a very handy tool in signal processing. The same can be further extended to two dimensional problems in image processing. The tool also emerged as a favourable method to provide solutions through the machine learning and other statistical techniques. In this paper, an attempt is made to employ the LRM to denoise the image which is subjected to guassian noise of certain variance. The performance analysis has been made in terms of calculated peak signal to noise ratio (PSNR).

     


  • Keywords


    Low Rank Matrix Method;Denoising; PSNR.

  • References


      [1] J. F. Cai, E. J. Cand_es, and Z. Shen. A Singular Value Thresholding Algorithm for Matrix Completion. SIAM J. Opt., 20:1956, 2010.https://doi.org/10.1137/080738970.

      [2] E. J. Cand_es and Y. Plan. Matrix completion with noise.Proc. IEEE, 98(6):925{936, 2010.

      [3] W. Dai and O. Milenkovic. SET: an algorithm for consistent matrix completion. 2009. Availableat http://arxiv.org/abs/0909.2705.

      [4] Ankit Parekh and Ivan W. Selesnick, “Enhanced Low-Rank Matrix Approximation”, IEEE SIGNALPROCESSING LETTERS, VOL. 23, NO. 4, APRIL2016.

      [5] W. He, Y. Ding, Y. Zi, and I. W. Selesnick, “Sparsity-based algorithm for detecting faults in rotating machines,”Mech. Syst. Signal Process., vol. 72, pp. 46–64, 2016.https://doi.org/10.1016/j.ymssp.2015.11.027.

      [6] H. Hu, J. Froment, and Q. Liu, “Patch-based low-rankminimization for image denoising,” [Online]. Available:http://arxiv.org/abs/1506.08353, 2015, pp. 1–4.

      [7] B. Huang, C. Mu, D. Goldfarb, and J. Wright, “Provablemodels for robust low-rank tensor completion,” Pac. J.Optim., vol. 11, no. 2, pp. 339–364, 2015.


 

HTML

View

Download

Article ID: 17569
 
DOI: 10.14419/ijet.v7i4.17569




Copyright © 2012-2015 Science Publishing Corporation Inc. All rights reserved.