Document Understanding - 2022.5 Community Preview

Greetings, UiPath Community! :dancer:

It’s been a while since we shared our updates - but now we’re ready to share our latest & greatest!

To improve the life of users validating the extracted data from documents, we have continued the series of improvements on the Validation Station. In this sense, we have introduced:

OCR & Extraction Confidence Scores in the Validation Station
With the latest update, the Validation Station provides the ability of users to see both OCR (for the engines reporting this property), as well as extraction confidence scores - and for you to be able to look at the most relevant extractions (where the confidence score is not high), we have introduced a filtering option, which groups the extraction results into 2 categories: above & beyond a minimum configured confidence level. If you modify an Extracted Value, its extraction confidence score will be set to 100% (while keeping the old value, available on mouse click) - so that the “requires attention” group of fields will not contain it anymore & you can focus on the remaining items.
However, please be careful when working with the confidence score - it should be used for guidance only and not treated as 100% accurate :slight_smile:
a0b18e9-22.5-VS31

The changes are available in the Validation Station for the Present Validation Station Activity, soon to be available in Action Center as well as (when using the Create Document Validation Action).

Upgrading the Omnipage Library
We have upgraded the Omnipage Library to its latest version - in this way, you are now able to benefit of the detection of hand-printed text (where the characters don’t have a connection). The ideal size for a document is between 25-45 pixels.

Migrating to .net 5 cross-platform
We have migrated the PDF & the Vision packages to .net5 cross-platform, so that they can be used from various operating systems :dancer:

The usual - stabilization & bug fixing
We have worked on small improvements of the UX of the Template Manager, allowing you to easier manage & remove unused Document Types.

15 Likes

Amazing… Was waiting for that filter option… Great stuff everyone!!! :slight_smile:

Great updates, as always, team!

Looking forward to trying the new version :grinning:.

1 Like

Was eagerly waiting for this feature. I understand we should be able to define the threshold value (80% or 90%) to be used from the Present Validation Station activity. :+1: