MAD-based Estimation of the Noise Level in the Contourlet Domain
- 12 Downloads
Noise-level estimation remains one of the most critical issues related to the contourlet-based approaches. In this paper, an investigation of an effective solution is directed from any redundant contourlet expansion. This is going to be addressed for the first time in that domain. In this proposition, the noise level is estimated as the median absolute value (MAD) of the finest multi-scale coefficients, calibrated by three correction parameters. This is done according to some visual classification of the natural images. The present estimator provides a better compromise between the image and the contourlet expansion nature, which makes the estimation results more accurate for a wide range of natural images, when compared to the best state-of-the-art methods. Therefore, it is extensively recommended for most of the contourlet-based image applications (Thresholding, filtering, etc.) thanks to its accuracy, simplicity, and rapidity.
Keywords:contourlet directional filter bank MAD noise-level estimation thresholding wavelet
The authors would like to thank Dr. Victoriya V. Abramova for providing the basis and the elements of the image classification adopted in this paper.
CONFLICT OF INTEREST
The authors declare that they have no conflicts of interest.
- 1.Yoon, B.J. and Vaidyanathan, P.P., Wavelet based denoising by customized thresholding, ICASSP, 2004, pp. II-925–II-928.Google Scholar
- 3.Arivazhagan, S., et al., A new hybrid image restoration method based on fusion of spatial and transform domain methods, Int. Conf. on Recent Adv. Comput. and Software, 2012, pp. 48–53.Google Scholar
- 6.Liu, C., et al., Noise estimation from a single image, Proc. IEEE Conf. Computer Vision and Pattern Recognition, 2006, pp. 901–908.Google Scholar
- 8.Liu, X., et al., Noise level estimation using weak textured patches of a single noisy image, 19th IEEE Intern. Conf. Image Process, 2012, pp. 665–668.Google Scholar
- 12.Chen, G., et al., An efficient statistical method for image noise level estimation, Proc. IEEE Intern. Conf. Computer Vision, 2015, pp. 477–485.Google Scholar
- 13.Das, A.K. and Chandle, J.O., Accurate noise level estimation through singular values and linear regression, 2nd IEEE Int. Conf. on Eng. & Tech. (ICETECH), 2016, pp. 1132–1135.Google Scholar
- 15.Kingsbury, N.G., The dual-tree complex wavelet transform: A new technique for shift invariance and directional filters, IEEE Digit. Sig. Proc. Workshop, Bryce Canyon, 1998, pp. 120–131.Google Scholar
- 16.Candes, E.J. and Donoho, D.L., Curvelets, multi-resolution representation, and scaling laws, International Symposium on Optical Science and Technology, International Society for Optics and Photonics, 2000, pp. 1–12.Google Scholar
- 18.Do, M.N., Directional multi-resolution image representation, Ph.D. Dissertation, Lausanne: Swiss Federal Institute of Technology, 2001.Google Scholar
- 19.Lu, Y. and Do, M.N., CRISP-contourlet: A critically sampled directional multi-resolution image representation, SPIE Conf. Wavelet Apps. Signal Image Process, San Diego, 2003.Google Scholar
- 20.Lu, Y. and Do, M.N., A new contourlet transform with sharp frequency localization, Proc. of the International Conference on Image Processing (ICIP '06), Atlanta, 2006, pp. 1629–1632.Google Scholar
- 24.Li, C.J., et al., Image denoising algorithm based on non-subsampled contourlet transform and bilateral filtering, International Conference on Computer Information Systems and Industrial Applications, 2015, pp. 666–669. https://doi.org/10.2991/cisia-15.2015.182.
- 27.Zhao, Q., et al., Mixed image denoising method of non-local means and adaptive Bayesian threshold estimation in NSCT domain, 3rd IEEE Int. Conf. on Comp. Sci. & Inf. Tech, 2010, pp. 636–639.Google Scholar
- 28.Zhou, Z., et al., Contourlet-based image denoising algorithm using adaptive windows, 4th IEEE Int. Conf. on Ind. Elec. & Apps, 2009, pp. 3654–3657.Google Scholar
- 35.A Matlab Contourlet Toolbox. https://www.ifp.uiuc.edu/~minhdo/software.Google Scholar