There is no need to extract
characters, register a dictionary, or adjust any other setting
configurations required for optical character recognition with
conventional vision systems. Just outline the text to identify it.
Furthermore, stable reading is possible even if the conditions of
the text’s shading, thickness, and size change.