Azure Document Intelligence correction

Kola, Sarath Babu 5

Hi, Currently we were using ABBYY/KOFAX for OCR. And both has a verification stage where user can correct the ocred text and the system will automatically learn and train itself. we want to replace with Azure DI and as we understand we cannot do the verification. Any idea how to achieve this functionality. Thanks, Sarath.

santoshkc 4,425 Reputation points Microsoft Vendor

2024-02-28T07:32:51.59+00:00

Hi @Kola, Sarath Babu,

Following up to see if the below answer was helpful. 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.

1 answer

santoshkc 4,425 Reputation points Microsoft Vendor

2024-02-27T07:32:32.0033333+00:00

Hi @Kola, Sarath Babu,

Thank you for reaching out to the Microsoft Q&A forum!

I understand that you are looking to replace ABBYY/KOFAX OCR with Azure DI, but you are concerned about the verification stage. The OCR model that can extract text from images and documents. However, it does not provide a verification stage where users can correct the OCR text and train the system.

One possible solution is to use Azure Document intelligence. The DI is a cloud-based service that uses machine learning to extract text, key-value pairs, and tables from documents. It also provides a labeling tool that allows you to label and train your own custom models. You can use Azure DI to extract text from your documents and then use a custom application to display the extracted text for verification.

Hope this helps. Do let us know if you any further queries.

If this answers your query, do click Accept Answer and Yes for was this answer helpful.
Please sign in to rate this answer.
Kola, Sarath Babu 5 Reputation points

2024-02-29T13:11:26.7366667+00:00

Thanks for your response. After correcting from custom application will Azure DI takes corrected data and train itself, this is what ABBYY/KOFAX does. they both learn on their own. Thanks, Sarath.

santoshkc 4,425 Reputation points Microsoft Vendor

2024-02-29T13:37:25.3366667+00:00

Hi @Kola, Sarath Babu,

Thanks for your follow-up question.

Yes, Azure DI can take the corrected data and use it to train itself. The Labeling Tool in Azure DI allows users to manually correct OCR text and train the system to improve future OCR results. The system uses this corrected data to learn and improve its OCR accuracy over time.

So, Azure DI can learn on its own and improve its OCR accuracy with the help of corrected data. For more info, please look into this complete Azure DI documentation.

I hope this answers your question. Let me know if you have any further questions or concerns.

Please do click Accept Answer and Yes for was thid and earlier provided answer was helpful. And, if you have any further query do let us know.

paul holden 25 Reputation points

2024-04-24T10:53:35.16+00:00

Can you show exactly what correction tool you are referring to as I cannot find it in the user interface for the custom documents in Azure DI. I have also referred to the Azure DI documentation and can find no mention of it.

paul holden 25 Reputation points

2024-04-24T11:06:15.21+00:00

Hi, what tool allows for the correction of OCR data in Azure DI?

santoshkc 4,425 Reputation points Microsoft Vendor

2024-04-24T12:04:34.3266667+00:00

Hi @paul holden,

I would recommend creating a new thread on the same forum with as much details about your issue as possible. That would make sure that your issue has better visibility in the community.
Sign in to comment