function [sigma, mu] = statsCluster(pixels, mask)% STATSCLUSTER - calculates the mean and STD of the given cluster%%  [SIGMA, MU] = STATSCLUSTER(PIXELS,MASK)%%  Input:%       PIXELS - MxNx3 image, where each pixel is a point in a %                perceptually unif color space (e.g. LAB). note that this image%                should have been masked (i.e. contain only those pixels in this cluster)%      MASK - MxN binary image that's a mask for this cluster%%  Output: %       SIGMA - standard deviation of this cluster. the STD calculated is%               the std of distance from the mean %       MU - 1x3 vector containing the mean L, A, and B values of this cluster%% Jeff Walters & Angi Chau% Feb 2003% first find the mean of the cluster (and we have to make sure to ignore% all the pixels not in this cluster)indices = find(mask~=0);Lonly = pixels(:,:,1);Aonly = pixels(:,:,2);Bonly = pixels(:,:,3);mu = [mean(Lonly(indices)) mean(Aonly(indices)) mean(Bonly(indices))];%if (numPixels > 0)    sd = sqdist(pixels, mu).*mask;    %sigma2 = sum(sigma2(1:end)); % sum of all distances    %sigma = sqrt(sigma2/numPixels);  % normalize & sqrt to get std        % now we want to find the std of the distances    sigma=std(sqrt(sd(indices)));    %sigma=std(sd(1:end));%end