The Matlab functions in this directory solve the deconvolution problem in the
paper D. Krishnan, R. Fergus: "Fast Image Deconvolution using
Hyper-Laplacian Priors", Proceedings of NIPS 2009. 

The code has been tested on Matlab 7.9 (R2009b). Our paper and the
code are related to the work and code of Wang et. al.:

Y. Wang, J. Yang, W. Yin and Y. Zhang, "A New Alternating Minimization
Algorithm for Total Variation Image Reconstruction", SIAM Journal on
Imaging Sciences, 1(3): 248:272, 2008, and their FTVd code.
  
The files in this directory are:

test_fast_deconv.m: A test script that reads in an image, adds noise
and blur to it and calls the deblurring function. The parameter values
lambda and alpha are set here. Noise level can be controlled as
well. Run this script in Matlab as "test_fast_deconv".

fast_deconv.m: The main non-blind deconvolution function that takes in
the blurry/noisy image, blur kernel, lambda and alpha values and
outputs a deblurred image. The parameters for the alternating
minimization scheme are set in this function.

snr.m: Computes Signal-to-Noise ratio between an input image and a
reference image. 

solve_image.m: Either computes a look-up-table to solve the w
sub-problem, or uses the existing look-up-table. The LUT is stored as
a persistent variable. There are a number of functions in this file to
compute the entries of the LUT for different values of alpha.

We use the MeX function pointOp from Eero Simoncelli's
matlabPyrTools. Please download this from:
www.cns.nyu.edu/~lcv/software.php and compile the pointOp MeX
function.  The function should be
visible from the directory where test_fast_deconv is run. If it is not
visible, fast_deconv will revert to using the much slower Matlab
function interp1 for LUT interpolation/extrapolation.

Update: January 1, 2010: A serious bug was found in the previously
uploaded version of the code from December 2009 - the test function
should output a deblurred clocktower image with SNR about 18dB (earlier, 
the bug cause the output blurred data to be 15dB).

Copyright (C) 2009: D. Krishnan, R. Fergus. 
Email: {dilip,fergus}@cs.nyu.edu

