0

I am starting out with medical images and I have managed to remove the entire background image. Here is an example image after I have performed the processing steps so far:

enter image description here

My goal is to separate the dotted lines from the rest of the artifacts my current processing steps haven't yet removed. Could someone point me in a good direction to completely separate these dotted lines from the rest of the image? I don't need anyone to write code for me (although you can for others if you like), all I need is what topics/links I should read up on. I should note the dotted lines come in various colors and are all either oval or linear in shape.

 Apparently my question is "too broad", so let me clarify.  How could I separate the dotted lines from non dotted lines in an image such as the one given here?
Beaker
  • 2,804
  • 5
  • 33
  • 54
  • OpenCV could be your choice (edge detection, primarily): https://stackoverflow.com/questions/34389384/improve-contour-detection-with-opencv-python – a small orange May 04 '19 at 22:47
  • 1
    Just some thoughts... First find all the "blobs" -- grayscale, very low threshold (basicaly black vs non-black), connected components or contours. Filter by size (drop anything too big). The dots all look quite similar, so maybe some simple pattern/template match (on grayscale or intensity, ignoring color) to filter out just the dots. Partition the dots based on color (maybe sum of absolute differences). Then try to fit a line/ellipse to each group. Eliminate outliers. More examples of what you could possibly get as input would help. May get tricky with multiple lines/ellipses of same color. – Dan Mašek May 04 '19 at 22:48
  • You can try image subtraction. – spaceman May 04 '19 at 23:16

0 Answers0