Extraction of MICR code, showing unacceptable results if the document is in the Japanese/Chinese/Korean language.
MICR fragment due to applied parameters constraints of MICR symbols which are significantly far from Normal Latin text type but closer to Japanese/Chinese/Korean symbols.
Taking into consideration that Japanese/Chinese/Korean languages and MICR are interfering, the best choice here will be using Region and Block methods to create a rectangle around the MICR code which will be processed outside the image.
As well as, we need to apply the language setting "English" and "TT_MICR" respectable to your code type to the rectangle in order to bypass the issue with languages.
As well as, it is important to use:
That would allow you to extract MICR correctly.