Is it possible to use classifier to differentiate annex pages from invoice pages in Vantage?
There are three different types of classification available in ABBYY Vantage.
It is not possible to use Classification Skill to specify certain invoice pages as Annexes. The Classification Skill works with the whole documents, not separate pages.
Classify Activity in Process Skill
It is possible to setup document pages reassembling based on classification results of the Classify Activity. Note that Assemble activity should be used to reassemble documents from pages, and the Assemble Activity works on per-page basis.
However, after the documents are reassembled, there will be new documents created from the original document, and each will be processed separately. In such conditions field extraction can be improved and training simplified by using only pages containing valuable data and splitting other pages into Annex document.
With such approach it is not possible to setup the original document output to PDF or JSON after the document pages are reassembled into new documents. Then, in case the documents pages are assembled back into the original state previous extraction results will be lost unless the results are saved elsewhere with the Custom activity script. The field extraction results will not be accessible with Public API methods.
Classification Activity in Document Skill in Vantage Advanced Designer
Classification Activity in Document Skill created with Vantage Advanced Designer can be used to classify documents to then select which Extraction Rules activity should be used for the document based on the Classification results. Still, the activity works with the whole documents, not certain pages.
It would be more appropriate to properly setup Extraction Rules activity in addition or separately from Classification Activity in case of creating Document Skill in Vantage Advanced Designer.
How to setup annex pages in ABBYY Vantage?
Please sign in to leave a comment.