Reg Expression for specific String

[Luke_Nguyen]

I'm trying to capture the check amount that is in a box in the following format $****6,897.71  .  I haven't been able to get the numeric value back in a consistent manner with the suspect level below 90.  Any ideas how I can improve this?  Use a regular expression?  Thanks.

[Steve_Reed · Accepted solution]

The simplest way to extract it might be to set the filtering to all character types, and then in the 'specific allowed characters' field, enter the values: 0123456789.  (note the period on the end)

This way the engine will attempt to read everything, then before assigning to the keyword, it will remove all characters except for the allowed ones (the numeric digits and the decimal point).

[Steve_Reed]

Which version of the OCR engine are you using? 

[Luke_Nguyen]

The listing in Add/Remove Program has Hyland OCR Engine for OnBase 16.5.

[Steve_Reed · Accepted solution]

It might be worth upgrading it to 16.6 (this will work with your existing OnBase version).  I don't know for sure that this is something that Nuance has addressed, but if the problem still occurs in the highest available version, then I can take it up with Nuance.

[Jeffrey_Will]

I've had issues with the * being read as a number too.  It may just bee a limitation on the OCR software.  Could a higher DPI setting help (i.e. 600 vs. 200)?

[Luke_Nguyen]

I upgraded to OCR 16.6 but got the same results.

I'll try upping the resolution just to see when the error clears, but I'm a little reluctant to go to higher resolutions due to file size.