Decipher - How to ignore everything after the first page

[isohrob]

Basically, we have trained Decipher on some documents that are single page and multi page, yet Decipher keeps going back to look at the second page even though all the values it needs are on the first. Is there a DFD flag or something I can set to ignore other pages? We are using Decipher for the purpose of data extraction, we are doing the verification ourselves, and so far this issue is one among a few that's been causing us issues.

[lookman]

Hi Isohrob.

The fields you've identified in your DFD are part of the header, not in a table. Correct?

If so, it sounds like Decipher is treating your multi-page document as separate documents because if Decipher finds a header field, it does not continue looking for it.

jack

[Daisy98]

To ignore everything after the first page, you can adjust your script to read only the first page of input. This ensures that processing stops after the initial content, saving time and resources in handling additional, unnecessary data.

[Athiban_Mahamathi]

Hi Isohrob,

If the fields belongs to the header, then please try the following misc-parm options for the fields which are not getting extracted from first page. 

• StrictPosition
• StrictPositionAnchorText
• StrictPositionBottomBoundary
• StrictPositionLeftBoundary
• StrictPositionRightBoundary
• StrictPositionTopBoundary

Refer : https://docs.blueprism.com/bundle/decipher-idp-2-3/page/user-guide/misc-parameters.htm

Regards,

Athiban

[Ben.Lyons1]

Hi,

Have you considered creating a classification model?

You could train page 1 as Doc Type A and the remaining pages as Doc Type B, the pages for each document type must be manually separated for training.

After training set Doc Type B as an attachment in the document type settings. This will instruct Decipher to ignore the pages for extraction.

Classification Model User Guide

Thanks