Back to the Homepage

Experimental results of the manuscript:

 

Weisheng Dong, Guangming Shi, Xin Li, Yi Ma, and Feng Huang , “Compressive sensing via nonlocal low-rank regularization”,

IEEE Trans. on Image Processing, vol. 23, no. 8, pp. 3618-3632, 2014.

 

The source code of this work can be downloaded from here: Matlab Codes

 

Note:

(1)   The TV method in [1] is labeled as “TV”;

(2)   The ReTV method in [2] is labeled as “ReTV”;

(3)   The BM3D-CS method in [3] is labeled as “BM3D”;

(4)   The MARX-PC method in [4] is labeled as “MARX”;

(5)   The proposed NLR-CS-baseline method is labeled as “NLR-bas”;

(6)   The proposed NLR-CS method is labeled as “NLR”;

Then, the reconstructed image by the method NLR-CS on image House is labeled as “NLR_House”. Other result images are labeled similarly.

 

Experiment 1: Natural images with random subsampling scheme

The reconstructed results on Lena256                    Download images

The reconstructed results on Monarch                    Download images

The reconstructed results on Barbara                    Download images

The reconstructed results on Boat                       Download images

The reconstructed results on C. Man                     Download images

The reconstructed results on Foreman                    Download images

The reconstructed results on House                      Download images

The reconstructed results on Parrots                     Download images

 

Experiment 2: Natural images with pseudo radial subsampling scheme

The reconstructed results on Lena256                    Download images

The reconstructed results on Monarch                    Download images

The reconstructed results on Barbara                    Download images

The reconstructed results on Boat                       Download images

The reconstructed results on C. Man                     Download images

The reconstructed results on Foreman                    Download images

The reconstructed results on House                      Download images

The reconstructed results on Parrots                     Download images

 

Note:

(1)   The CS-MRI method in [5] is labeled as “SparseMRI”;

(2)   The zero-filling method is labeled as “ZF”;

(3)   The proposed NLR-CS-baseline method is labeled as “NLR-bas”;

(4)   The proposed NLR-CS method is labeled as “NLR”;

Then, the reconstructed image by the method NLR-CS on Head MR image is labeled as “NLR_Head”. Other result images are labeled similarly.

 

The original MR images:

The original MR image Head                            Download image

The original MR image Brain                            Download image

 

Experiment 3:  MR images with 2D random sub-sampling mask

The deblurring results on Head                          Download images

The deblurring results on Brain                          Download images

 

Experiment 4:  MR images with pseudo random radial mask

The deblurring results on Head                          Download images

The deblurring results on Brain                          Download images

 

 

References:

[1]   Http://www.acm.caltech.edu./limagic

[2]   E. Candes, M. Wakin, and S. Boyd, “Enhancing sparsity by reweighted l1 minimization,” Journal of Fourier Analysis and Applications, vol. 14, no. 5, pp. 877-905, 2008.

[3]   K. Egiazarian, A. Foi, and V. Katkovnik, “Compressed sensing image reconstruction via recursive spatially adaptive filtering,” in IEEE International Conference on Image Processing, vol. 1, San Antonio, TX, USA, Sep. 2007.

[4]   X. Wu, W. Dong, X. Zhang, and G. Shi, “Model-assisted adaptive recovery of compressed sensing with imaging applications,” IEEE Trans. on Image Processing, vol. 21, no. 2, pp. 451-458, Feb. 2012.

[5]   M. Lustig, D. Donoho, and J. Pauly, “Sparse MRI: The application of compressed sensing for rapid MR imaging,” Magn. Reson. Med., vol. 58, no. 6, pp. 1182-1195, 2007.