I want to diagnose disease malaria from blood sample. And I had done almost. But, there is a problem with the overlapped rbc. How to extract overlapped rbc from blood smear image?

This is my code.

% Find image size
[m,n,numChannels] = size(im);

bw1=k_means(im*1.2);
bw1=im2bw(bw1);
bw2=bw1;

% Find Image Properties
reg_pro=regionprops(bw2,'all');

% Find small regions
for i=1:length(reg_pro)
    if reg_pro(i).Area>100000
        reg_pro(i).PixelIdxList=1;
    end
end

% Save each images seperately as png

bboxes = [];
fnNum = 1;
fprintf('\nSaving Individual RBC images...')
for i=1:length(reg_pro)
    k=reg_pro(i).BoundingBox;
   if k(3)>50 && k(4)>50 && k(4)/k(3)<2 && k(4)/k(3)>.6

          g=imcrop(im,[k]);
          imwrite(g,strcat(segmentedPath,num2str(fnNum),'.png'));
          imshow(g)
          bboxes = [bboxes ; k];
          %         if dbug
          %             imshow(g)
          %             pause(.1)
          %         end

          fnNum = fnNum + 1;
       end
  end

[Walter Roberson]

Sample images make it easier to reply.

[Image Analyst]

No idea, since you're not sharing your image. In the meantime, see

https://www.mathworks.com/help/images/marker-controlled-watershed-segmentation.html

https://blogs.mathworks.com/steve/2013/11/19/watershed-transform-question-from-tech-support/

http://www.mathworks.com/matlabcentral/answers/6200-tutorial-how-to-ask-a-question-on-answers-and-get-a-fast-answer