I would like to find the connected components in this image (the image is binary)
and for that, I use the OpenCV cv2.connectedComponentsWithStats algorithm. But the latter gives me a result that I can't understand. Indeed, if we take the first connected component, we find that there are several parts of this component that are not connected (see image, the component is in white).
Could you explain to me the logic behind this result?
PS 1: I tried with other connected component algorithms (such as skimage.mesure.label) and I got the same results!
PS 2: For comparaison, here's the second connected component (always in white)