Connect HomeMarketplaceComponentsExtract Semi Structured Document Activity

Extract Semi Structured Document Activity

Custom Activity

Custom Activity

Updated: October 30th, 2020

Published: July 17th, 2019

Downloads: 716

Language compatibility: Visual Basic

Community Support

Extract Semi Structured Document Activity

Custom Activity

Summary: 

A custom activity to query UiPath's machine learning models for semi-structured document data extraction

The Extract semi-structured data activity allows RPA developers to easily take advantage of UiPath's machine learning models for semi-structured documents processing. The models currently can analyze invoices and receipts, providing various information (total price, vendor address, list of items, etc.). By simply specifying the image to process and the address of the model to use, the activity returns the model's response already formatted for parsing or verification using UiPath's Validation Station.
A typical use case would be to retrieve the total and vendor name from a list of scanned receipts and input them in an expense managing software (e.g. Concur). An advanced version of this process could even check for confidence scores and trigger a different workflow which will present validation station to a human user.
There are currently two machine learning models that can be targeted by the activity:
  • Invoices: https://invoices.uipath.com
  • Receipts: https://receipts.uipath.com

Details

Benefits

Easily integrate any UiPath machine learning model in your automation with one activity
Directly generates data to be used with Validation Station and other IntelligentOCR activities
Easy to parse/troubleshoot output

Compatibility

Studio 2018.4 +, IntelligentOCR activities 2.0.0+

Dependencies

System.Collections.Immutable >=1.5.0
Nito.AsyncEx.Coordination >= 5.0.0
UiPath.DocumentProcessing.Contracts >= 1.0.0
UiPath.PDF.Activities >= 3.2.2

Resources

- Extract semi structured data activity.pdf

Licensing

By clicking download you agree to the following license.