Clear Filters
Clear Filters

How can I remove the circles(different colours) from the dermoscopic image?

4 views (last 30 days)
My dataset contains a variety of noise. One noise is in the form of different shapes(circles) of different sizes and colors at different locations. How can I remove them? Please give me some guidelines or some hints/code from where to start. Please see the attached image. I will be really thankful
Nazia Hameed
Nazia Hameed on 3 Oct 2017
Edited: Nazia Hameed on 3 Oct 2017
ok. Thank you, I will have a look. I tried color segmentation and getting initial results. Thank so much.
Nazia Hameed
Nazia Hameed on 3 Oct 2017
I have tried and I generated a mask. Can you plz tell me how can I fill the holes? I have tried your code available on But its not working. I have also tried imfill function. for filling holes I am writing the below code
smallestAcceptableHoleSize = 4000; % in pixels.
fillimage2 = bwareaopen(fillimage, smallestAcceptableHoleSize);
subplot(4, 4, 13);
imshow(fillimage2, []),title('Fill Image');

Sign in to comment.

Answers (2)

Selva Karna
Selva Karna on 7 Oct 2017
first separate region based object, after that apply contour , further apply binary selection .

Image Analyst
Image Analyst on 7 Oct 2017
What I'd do is to threshold the saturation channel since the colored circles have much higher saturation than the skin. Attached is a demo.
  1 Comment
Nazia Hameed
Nazia Hameed on 9 Oct 2017
First of all Thank you so much for your time you are spending to help me. I am really really thankful to you and may you have the best of all the things. I tried your code but this is also detecting the area of interest. All I want is to create a generic code/algo that can be applied on all the images (with circles/without circles) with different types of diseases. The benign cancer is really small but the malignant cancer are really large in size and also the colour is different. When I apply the one developed by you as well as by me it detects the region of interest too. Please see attached picture. is there any way I can detect using shape?
once again thank you so much.

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!