There is no need to perform extraction, adjust character width / height, register a dictionary, or configure any other settings required with conventional vision systems. Just outline the target text to identify it. Furthermore, stable reading is possible even if the shading, thickness or size of your text changes.