Extract values for a set of fields in a document using Azure Document Intelligence

Question

Is it possible to get the Document Intelligence service to extract values for a defined set of fields?

Seeing all the document's fields, tables and other data structures is great when deciding which data points I want extracted from the document. Once I know which data points I want extracted from a document, it would be great to only have those data points extracted.

I would expect it to be quicker to process the document with the sub-set of fields. Will it be quicker?

If this feature is supported, can it be added to the document intelligence library?

https://learn.microsoft.com/en-us/dotnet/api/overview/azure/ai.documentintelligence-readme?view=azure-dotnet-preview

Answer 1

@Bijaya Rai You will have to train a custom extraction model to extract fields that are of interest to you. The trained model will then be used to extract fields from documents similar to your trained documents. Please go through the tutorial to build a custom extraction model through the DI studio and it should give you an idea on what is possible with a custom model when you train the model to extract only certain fields.

Once you are satisfied with the model you can call it programmatically and use it in your applications to extract the required fields of documents. Thanks!!

If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.