This software is the MATLAB code for the image deconvolution algorithm in the paper
Chao Jia and Brian L. Evans "Patch-based image deconvolution via joint modeling of sparse priors", 
submitted to IEEE International Conference on Image Processing (ICIP) 2011.

The code has been tested in MATLAB 7.9.0 (R2009b). The test images are included in "\test images"
directory and can be used to repeat the experiment result in the paper. The original, blurred and 
deconvolved image will appear in "\result" directory after running the demo.

The software includes files as follows:

joint_sparse_deconv_demo.m
	The demo function of the software. The parameters are specified in the function and can be tuned
	freely by users. Users can also change the test image, the convolution kernel and the variance 
	of additive noise for different test settings.
	
sparse_dec.m
	The main function of image deconvolution using the joint model proposed in the above mentioned
	paper.
	
sparse_solve_x.m
	A function used to solve x sub-problem in the paper.
	
imtocols.m
	A function for extracting disjoint patches from an image.
	
dictionary.mat
	A learned dictionary for sparse representation of image patches, with size 12*12. The dictionary 
	is learned based on an online learning algorithm in J. Mairal, F. Bach, J. Ponce and G. Saprio, 
	"Online dictionary learning for sparse coding", ICML 2009. The patches for learning are extracted
	from Berkeley segmentation database.
	
The software also needs several functions in some released softwares written by others. These functions 
(with copyrights) are included in the "\tools" directory:

a function 'psnr' from the Image Halftoning Toolbox from http://users.ece.utexas.edu/~bevans/projects/halftoning/index.html
a function 'ssim_index' to compute the Structural Similarity Measure by Z. Wang from http://www.ece.uwaterloo.ca/~z70wang/research/ssim/
a function 'mexLasso' to solve the Lasso problem from http://www.di.ens.fr/willow/SPAMS/
a function 'solve_image' to solve L_p norm shrinkage from http://cs.nyu.edu/~dilip/wordpress/?page_id=122

Please also see the COPYRIGHT.txt file for copyright.
If you have any problem with this software please contact Chao Jia by email: kurtjc@gmail.com.