Standardizing Table Schema from Prebuilt - Invoices Using Azure Document Intelligence

Vikas M 60 Reputation points
2024-05-24T05:58:19.7566667+00:00

I am using the Prebuilt Invoice Model from Azure Document Intelligence. While the model works well overall, I am facing an issue with the table schema. The invoices I receive come from various sources, each with a different table structure.

My goal is to map the extracted values to a fixed table schema and display it in my program with fixed table columns and headers.

Challenges :

  • Invoices have varying table structures.
  • Need to standardize these tables into a uniform format and then map it to my program table

I aim to create a consistent table format from various invoice structures, making the data easier to manage and display in my application.

Azure AI Document Intelligence
Azure AI Document Intelligence
An Azure service that turns documents into usable data. Previously known as Azure Form Recognizer.
1,505 questions
0 comments
Accepted answer
  1. romungi-MSFT 43,656 Reputation points Microsoft Employee
    2024-05-24T08:28:28.89+00:00

    @Vikas M With pre built models you cannot modify the model behavior to create separate structure in the resulting output response. You can however have a query field extraction with certain prebuilt models.

    Another option is to use custom model to train the model with all the required formats and then use the extracted result in your client application. You also have an option to create table field or tag to read certain data of a larger table and save it in this tag. But if the format varies frequently you will have to train several models to extract all the relevant data.

    If the current pre-built model works well, try using the output of this response in your client application to build the required table format. If your forms use a uniform header then it should be easy for you to use this data from the response and build on top of it in your client application.

    0 comments

0 additional answers

Sort by: Most helpful
Most helpful Newest Oldest