Two possible approaches come to mind here:
- The simplest approach is to apply the fuzzy edge detection and the superpixel clustering separately to the image, and then use the superpixel labels to create a mask to perform whatever processing you wish on fuzzy edges that lie within each of your regions of interest.
- The second approach is to use the superpixel labels to create masks for each of the regions of interest, and then pass the entire image to a separate function to perform fuzzy edge inspection after applying the mask to it. I am not sure what benefits this approach will give you, but the fact that it performs fuzzy edge inspection once for each of your regions of interest means this is by far the longer-running of the two appraoaches.
The second approach could look something like this:
numRegions = 20;
[L,N] = superpixels(Irgb,numRegions);
edgesForRegions = zeros([numRegions size(Irgb,1) size(Irgb,2)]);
for i = 1:numRegions
mask = L;
mask(mask ~= i) = 0;
mask(mask == i) = 1;
maskedRgbImage = bsxfun(@times, Irgb, cast(mask, 'like', Irgb));
edgesForRegions(i,:,:) = getEdges(maskedRgbImage);
end
function Ieval = getEdges(Irgb)
% code from https://www.mathworks.com/help/fuzzy/examples/fuzzy-logic-image-processing.html?
end
