Google Vision OCR recognizes for hexadecimal numbers with mistakes very often (the accuracy is about 60%). For example when I try to recognize a scanned image with muber &qu
