Many of the document image binarization algorithms assume the black texts are on a white background or vice versa. My question is what I can do if white and black texts coexist in one document image. Any suggestion will be welcome.
Detect the black regions and invert them.