Which is the most efficient approach to extract 'text' region from a given scene/image?

Similar questions and discussions