we first define the text area by various computer vision methodologies
then apply OCR to extract the text. In this process, each individual case is unique in the fonts used, and language improvements may be required to get the most accurate character recognition output.