MATLAB: Is there is any difference between otsu & graythresh?

image processingimage segmentationMATLAB

i am trying to make image segmentation using otsu's method but i am confused about the difference between using that function (IDX = otsu()) or using graythresh(I) ,Regarding that the first function didn't work and always that error appears (??? Undefined function or method 'otsu' for input arguments of type'double'.)…

Best Answer

You may find Otsu doesn't work well. It seems to work best for high contrast images with well separated bimodal histograms. If you have a skewed unimodal histogram, it often doesn't pick what you'd pick if you did it manually (with my thresholding app: http://www.mathworks.com/matlabcentral/fileexchange/29372-thresholding-an-image). Personally I like the triangle method better - it seems to work with more images that are continuous gray scale.

Related Solutions

MATLAB: Fill holes in binary Image

You could try imclose() followed by imfill()

se = strel('disk', 20, 0);
mask = imclose(mask, se);
mask = imfill(mask, 'holes');
imshow(mask);

or you could try an active contour, like the attached demo.

There are fancier ways if you want to make sure the outer boundary doesn't change, but it's more work and I don't have time for that now. See if one of the two methods I gave above works OK for you.

MATLAB: What’s the difference between GrayThresh and OtsuThresh

The only significant difference is that graythresh wants the image as input, whereas otsuthresh wants the image histogram instead. You can look at the actual code of both

edit graythresh
edit otsuthresh

and you'll see that all graythresh does is to compute the image histogram with 256 bins before calling otsuthresh.

Best Answer

Related Solutions

MATLAB: Fill holes in binary Image

MATLAB: What’s the difference between GrayThresh and OtsuThresh

Related Question