%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% >>>> IMAGEBOX >>>> JX >>>> UCLA >>>>
%%
%% image toolbox
%% MATLAB file
%% 
%% snr.m
%% compute signal-to-noise-ratio (SNR) of a noisy signal/image
%%
%% function s = snr(noisydata, original)
%%
%% input:  noisydata: noisy data
%%         original:  clean data
%% output: s: SNR value
%% example: s = snr(f, eu);
%%
%% created:       11/23/2003
%% last modified: 12/03/2005
%% author:        jjxu@ucla
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

function s = snr(noisydata, original)

[m,n] = size(noisydata);

mean_original = mean(mean(original));
tmp           = original - mean_original;
var_original  = sum(sum(tmp.*tmp));

noise      = noisydata - original;
mean_noise = mean(mean(noise));
tmp        = noise - mean_noise;
var_noise  = sum(sum(tmp.*tmp));

if var_noise == 0
    s = 999.99; %% INF. clean image
else
    s = 10 * log10(var_original / var_noise);
end

return