gocr -p ./micrdb/ -m 256 -m 130 -m 2 -s 80 -d 20 -C 0123456789CcAa -i /tmp/.1173296049117329605632_MICR.png
||" = CCc
|: = Aa
see man gocr for more
You want to train the database against the MICR extractions that the check module generates
---
NOTES
All MICR fonts must meet ANSI standard X9.27-1995.
http://www.micrfonts.com/what.htm
the MICR has only 14 characters
0-9
Transit
Ammount
On-Us
dash
each char occupies exactly 1/8th of an inch although each char itself occupies one of 5 possible widths
codes are ordered as:
ROUTING_NUMBER CHECK_NUMBER BANK_ACCOUNT_NUMBER
---
How is the MICR data on U.S. check organized? There are five major fields on a MICR line : 1. Auxiliary On Us - contains check number if present; bracketed by on us symbols 2. EPC - one character located to the left of the transit field if present 3. Transit - 9 digits including check digit 4. On Us - variable length 19 digits max; between transit and amount fields 5. Amount - 10 digits zero filled; bracketed by two amount symbols
Two of the five major fields are broken into multiple fields:
Transit (9 digits)
6. Routing Number (digits 1-4)
7. Bank Number (digits 5-8)
8. Check Digit (1 digit)
On Us
9. Account Number - Variable length; always followed by the On Us symbol
10. Check number - Located in various places in the on us field
11. TPC - max 6 characters; Located to right of account number