In a program of mine, some contours are detected in a binary image, then edited via a some manual brush tool, and then updated with another detect contour run. I noticed that during those iterations my objects get bigger and bigger, and I think it happens when I draw the contours. For instance, circles increase 1 pixel in diameter every time. This happens with all three functions I have tried (What is the difference between cv2.polylines, cv2.fillPoly, and cv2.drawContours?). It seems all of them add a border or boundary line.
How can I draw only the inside of a contour or polygon, and nothing else? Do I have to edit the coordinates directly to achieve that?