untitledLei Zhang¡¯s homepage


Experimental results of paper

 

¡°Two-stage Image Denoising by Principal Component Analysis with Local Pixel Grouping¡±

by Lei Zhang, Weisheng Dong, David Zhang and Guangming Shi

 

Pattern Recognition, vol. 43, issue 4, pp. 1531-1549, April 2010.

 

Paper:                   download here.

Matlab code:          download here.

 

The Matlab source code of our LPG-PCA denoising algorithm is optimized. Now it requires only 1~2 mins to denoise an image of 256 by 256!

 

We label

 

the denoising method [1] as ¡°pbshrink¡±;

the denoising method [2] as ¡°mixscale¡±;

the denoising method [3] as ¡°ksvd¡±;

the denoising method [4] as ¡°bm3d¡±;

and the proposed LPG-PCA-based denoising method as ¡°lpgpca¡±.

 

With these labels, the denoising result by using the method ¡°lpgpca¡± of noisy image Lena with =20 is saved as ¡°lena20_lpgpca.tif¡±, and similarly for other methods.

 

Experiment 1 -- Lena

download images

Experiment 2-- Cameraman

download images

Experiment 3 ¨C House

download images

Experiment 4 ¨C Barbara

download images

Experiment 5 ¨C Peppers

download images

Experiment 6 ¨C Paint

download images

Experiment 7 ¨C Monarch

download images

Experiment 8 ¨C Bloodcell

download images

Experiment 9 ¨C Tower (color)

download images

Experiment 10 ¨C Parrot (color)

download images

 

             

The PSNR (dB) and SSIM results of the denoised images at different noise levels and by different schemes. The value in the parenthesis is the SSIM [5] measurement.

 

 

Reference

 

[1]      A. Pizurica and W. Philips, ¡°Estimating the probability of the presence of a signal of interest in multiresolution single- and multiband image denoising,¡± IEEE Trans. on Image Processing, vol. 15, no. 3, pp. 654-665, March 2006.

[2]      J. Portilla, V. Strela, M. J.  Wainwright and E. P. Simoncelli, ¡°Image denoising using scale mixtures of Gaussians in the wavelet domain,¡± IEEE Trans. on Image Processing, vol. 12, no. 11, pp. 1338 ¨C 1351, Nov. 2003.

[3]      M. Aharon, M. Elad, and A.M. Bruckstein, ¡°The K-SVD: An algorithm for designing of overcomplete dictionaries for sparse representation,¡± IEEE Trans. On Signal Processing, vol. 54, no. 11, pp. 4311-4322, November 2006.

[4]      K. Dabov, A. Foi, V. Katkovnik, and K. Egiazarian, ¡°Image denoising by sparse 3D transform-domain collaborative filtering,¡± IEEE Trans. Image Processing, vol. 16, no. 8, pp. 2080-2095, August 2007.

[5]      Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, ¡°Image quality assessment: From error visibility to structural similarity,¡± IEEE Transactions on Image Processing, vol. 13, no. 4, Apr. 2004.