Question

Quickfield Session using ZoneOCR not consistent when Page Removal runs after it.

Quick Fields

Updated June 24, 2014

asked on April 6, 2014 • Show version history

Using Quickfields 9.x, I setup a session that IDs a document via a split sheet with a single barcode on it. Once it IDs a document it looks at OCR Zone on the second page to pull a P.O. number. The session was working great and was reading very consistently. That is until I added a Page Removal process after the Zone OCR. It was setup to remove the split sheet from 1st page of the document.

Once I did that the consistently took a really bad hit. Out of 8 test documents I was using 4 out of the 8 would not longer return good zone results.

I then removed the Page Removal process and it started reading 8 for 8 again. I added it back again and the read rate went back down to 4 of 8. I removed it again and it went back to 8 for 8. I kept doing this and it would always follow that pattern.

So I added Page Removal to the post process to get around it, which works fine for them. I am just wondering why this is happening and if this is a know issue?

John

1 0

Answer

SELECTED ANSWER

replied on April 7, 2014

Hey John,

I've done a bit of research to try and work out what was going on.

Basically it looks like the Page Removal was impacting the total number of pages. This meant that the Zone OCR was being applied to the wrong page.

As you noted, you can put the Page Removal in Post Processing, this way the total number of pages is reduced, but its after the Zone OCR, so it wont affect it.

If you need the Page Removal in the Page Processing, you could also try applying the Zone OCR to one page earlier. This should also fix the problem.

0 0

Replies

replied on April 7, 2014

Nathan it is my understanding that it works from the top down when doing the page processing. So it should have Zoned the correct area and then removed the 1st page.

If your theory is correct that means there is no order in how it runs the processes. If my understand is/was correct that means there is a bug.

What are your thoughts?

John

0 0

replied on April 7, 2014

The processes run in the order they are placed, top down as you mention.

However these processes only activate when meeting the "conditions" of being the right page, and they work on each page, one page at a time.

Page Removal is placed after the Zone OCR, but because it is targeted for Page 1, it will activate on Page 1 before Zone OCR activates on Page 2.

This is the design of Quick Fields, and so not technically a bug (everything is working as intended). Perhaps there is an argument that this design has weaknesses (for example in the case we have been looking at), but it may also be avoiding other problems too. A developer will be able to give you a better idea here.

0 0

replied on April 7, 2014

Nathan,

Very interesting... So if I understand the logic correctly it goes like this...

Example three page document.

Page 1:

Zone OCR - I am not looking at page 1 so I will ignore you.

Page Removal - I am looking at page 1 so I will run.

Moves to page 2.

Page 2:

Zone OCR - I am looking for page 2 so I will run, but now you are page 1, so go to the new page 2 which was page 3 and zone there.

Page Removal - I am looking for page 1 and since you were page 2 I will ignore you.

Moves to page 3.

Page 3:

Zone OCR - I am not looking for you even though your new number is page 2

Page Removal - I am looking for page 1 and since you were page 3 I will ignore you.

Am I on the correct track? Did I misunderstand your reply?

John

0 0

replied on April 7, 2014

No, that's not correct. Page 3 becomes page 2, so Zone OCR will run on that page. If you want it to run on the original page 2, then set it on page 1 and put Page Removal before Zone OCR.

0 0

replied on April 7, 2014

Miruna,

You are correct that is the best solution but what I am wondering is why does it use this logic? If it is indeed the best logic to use could there be some "warning" built into the Quickfields that tells the end user that it might not produce what is expected?

John

0 0

replied on April 7, 2014

It had to be designed this way because Quick Fields only keeps track of the number of pages in the scanned document, it has no concept of the number of pages in the original document. It won't know that page 2 in the original doc is now page 1 of the doc it has (for that matter it wouldn't know if you removed 1 or 2 or 10 pages until you got to what is now the first page - for ex, Page Removal set to page 1 followed by Page Removal by size). Coming up with any sort of mapping between the scanned pages and the existing pages in the document would be hard to maintain and even harder to represent to the user. So you need to set your Zone OCR process on page 1.

What makes the configuration slightly more complicated in this case is that users wanted to be able to read data from the original first page, remove the page and still have that data available as tokens. This is where the order of the processes comes in. If you have Zone OCR followed by Page Removal, both set to page 1, you get the Zone OCR data from the original page 1. If you flip the order, then Zone OCR only runs after Page Removal has completed its job, so it runs on what is now page 1, aka page 2 in the original document.

0 0

replied on April 7, 2014

Thats the idea. I will amend with what is a slightly more accurate interpretation.

Page 1:

Zone OCR - I am not looking at page 1 so I will ignore you.

Page Removal - I am looking at page 1 so I will run.

Page 1 is removed. Page 2 becomes Page 1. Page 3 becomes Page 2. There is no longer a Page 3.

Moves to page 2.

(this is the second page that QF looks at, but it is actually page 1 now)

Page 2:

Zone OCR - I am looking for page 2 so I will run, but now you are page 1, so go to the new page 2 which was page 3 and zone there.

(this was waiting for page 2, but as mentioned, this is called page 1 now. zone ocr does not run.)

Page Removal - I am looking for page 1 and since you were page 2 I will ignore you.

(this is page 1, and under those rules I would have expected page removal to activate again. however it does not, which implies something is built into page removal to only run on the real page 1, or there is something else that is going on)

Moves to page 3.

Page 3:

Zone OCR - I am not looking for you even though your new number is page 2

Page Removal - I am looking for page 1 and since you were page 3 I will ignore you.

(this page is now page 2. zone ocr activates on page 3 of the document, because it is now page 2!)

Its a confusing situation to explain, hopefully I have helped rather than hindered.

0 0

replied on April 7, 2014

Nathan that was very helpful, and very strange. I agree there must be a good reason why it works this way. So here is my real concern, if page removal is smart enough not to run again on each new page 1, but Zone OCR is not set to keep track of what page it was/is/changes to what other Quickfields plugins are inconsistent? For example I have other Quickfields clients that are having barcode misread issues. Those have all kinds of different plugs involved in an order that make sense. I wonder if any of those could be affecting me negatively.

Miruna has got the right idea though, put Page Removal before Zone OCR and that would firm everything up. I bet most people would still place their zone on page 2 though...

John

0 0

replied on June 24, 2014

FYI this is one of the reasons I have taken to putting page removal in workflow.

I also do it because I may need to validate a date field or some other type of field that has the potential to get "stuck" inside QF agent. This way if the date or other field doesn't get validated then it sits in a review folder for a user to manually fix it - with the needed cover page still there. Only once it's been validated does workflow remove page 1.

0 0

You are not allowed to follow up in this post.

Question

Question

Quickfield Session using ZoneOCR not consistent when Page Removal runs after it.

Answer

Replies

Sign in to reply to this post.