Question

Pattern Match when first line text is not always there.

Quick Fields

Updated February 2, 2016

asked on January 28, 2016 • Show version history

I have several years of time sheets that I need to Scan into laserfiche. My problem is that there are two variations of the time tickets within each pay period. I need a way to do first page Identification and to be able to consistently get Employee's name.

I have played with multi-value tokens and single line tokens.

This works for the time sheet directly above however fails for the first one. Looking for a solution that covers both. My OCR zone needs to be large because depending on how the time sheet was sent into office or scanned the data can move around a fair bit.

0 0

Replies

replied on January 28, 2016

My first suggestion would be to try changing the Identification Condition from

Token 1 equals EMPLOYEE NAME:

Token 1 contains EMPLOYEE NAME:

because there might be extra spaces that get captured that would make the text not match exactly.

However, to diagnose the issue further I'd need to know more information from the processing information/ output pane to determine what text is being captured by your Zone OCR in each case.

0 0

replied on February 1, 2016

I did try that and no change in results.

More research seems to indicate that when I scan in a time sheet followed by a no time sheet it is failing at this stage.

The non document is being attached the first page but the document fails and is placed on the left side in the document manager

0 0

replied on February 1, 2016

Can you show me configuration information for the Zone OCR process that creates Zone 1? And the output in the Processing Information pane for Zone 1?

I'm also unclear on why you need a Pattern Matching process here. Are you simply aiming to identify the page if it contains the words "EMPLOYEE NAME" in that Zone? If so, you can do the same thing with just the Zone OCR process.

0 0

replied on February 2, 2016

Thanks Tessa. I thought I needed to create a token for use in the Identification Condition area. I was able to delete my Token as you suggested.

I do have things working now with a 99% success rate. I also changed the advanced options by setting the single line option to false. Which seems also to have help with the identification

Part of my issue was that I was also running a pattern match to pick out first and last names from the time sheets plus there were other non time sheets mixed in. I was trying to do to much at once. Which compounded the errors.

0 0

You are not allowed to follow up in this post.

Question

Question

Pattern Match when first line text is not always there.

Replies

Sign in to reply to this post.