removing tags from breast thermal image
    4 views (last 30 days)
  
       Show older comments
    
need help to remove background and tags from breast thermal image, I have tried using threshold but it removed some important features I want to use later in classification of breast cancer.
this is the code I used in matlab
[fn,pth]=uigetfile('*.jpg')%%to open an image from any folder
filename=[pth  fn]
a=imread(filename); 
figure; imshow(a); 
bb=rgb2gray(a); 
figure; 
 imshow(bb);
bb = im2double(bb);              %convert to double
 thresh_level = graythresh(bb);  %find best threshold level
c = bb> thresh_level;%do thresholding
figure; imshow(c)
d = bwareafilt(c,1);
figure;imshow(d)
e = im2double(d).*bb;           %c is binary, so we need to convert it first
 figure;imshow(e,[])
0 Comments
Answers (1)
  Image Analyst
      
      
 on 14 Jan 2018
        It's better to just export the original thermal image without the annotation in there at all, rather than to try to fix up an annotated image. Can you get the original image? I would think so.
2 Comments
  Image Analyst
      
      
 on 15 Jan 2018
				The images can't be reliably processed in those areas because the data is missing. The best you can do is to estimate what the data should be. It looks like the annotations are all in fixed places. The easiest thing you can do is to make a binary image mask from those known locations, then use regionfill() to smear in values from just outside the boundaries.
See Also
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
