How can i detect homogeneous areas??
Show older comments
Hello, How can i detect homogeneous areas in an image with matlab?? thanks
Answers (1)
Image Analyst
on 2 Oct 2012
0 votes
I would try some spatial filters like stdfilt() and entropyfilt() to get images with high signal where there is high variance. Then threshold that to select dark areas, which are the "uniform" areas. What you do after that depends on what you want to do once you have the uniform areas.
7 Comments
Pamela Paolo
on 5 Oct 2012
Edited: Pamela Paolo
on 1 Nov 2012
Image Analyst
on 5 Oct 2012
OK. You already know how to upload images, so come back if you need help.
Image Analyst
on 7 Oct 2012
There are lots of papers on characterizing skin lesions. Go to section "20.13.3 Medical Applications -- Skin Cancer, Melanoma, Skin Lesions" at VisionBib. From the looks of your image, you could easily find it with just color segmentation, or maybe that combined with texture analysis. But those papers have already investigated this for decades and they will give you the best algorithms, probably way better than just saying to do texture analysis.
Pamela Paolo
on 9 Oct 2012
Edited: Pamela Paolo
on 1 Nov 2012
Image Analyst
on 9 Oct 2012
SO, have you totally given up on trying to understand? Or are you still working on it?
n ali
on 23 Mar 2015
I am doing image segmentation by using a matlab code, im getting the final output in the form of image size matrix sax 320*280 with quantized segemnted values, now i want to extract the different similar regions from it, can any one guide me abt any possible way , the regions are not homogeneous they varray in dimension and depends upon scene conditions
Image Analyst
on 23 Mar 2015
Post your image in a new question. I'm not sure what you're asking. You say you're already getting the segmented output, so what's left to do?
Categories
Find more on Texture Analysis in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!