Yes, you can do that multiple ways
- Get the two highest-spaced white pixels in the image, column numbers (left column and right column), number of rows (if luckily the same row, no problem, otherwise average (round) between two different rows , Which will be almost the same)
- Morphological operation, get the indices of furthest distance white pixels (horizantal & Vertical) and fill the region with white pixels.
- Thinning Algorithm (see bwmorph)
