There are many PNG images in the directory, some are in Chinese, some are in Japanese, and some are in Korean.
When using the OCR on a Japanese Windows system, Chinese and Korean images high error rate.
How can we increase the character set recognized by OCR?