AnalyzeExpense - Analyzing Invoices and Receipts

0

Hello there,

We are interested in using the recently announced Textract Invoice and Receipts API, but we want to know if you are planning any asynchronous options to process pdf directly without converting to image, just as we can do with ocr, tables or forms.

By the moment, the best the best solution we have thought of is to split pdf in n images and call the api for each one.
Is there any other better approach to do it?

Thanks,

질문됨 3년 전673회 조회
3개 답변
0

Hello,

Happy to hear that you are interested in Textract Invoice and Receipts API.
Document for invoice and receipts : https://docs.aws.amazon.com/textract/latest/dg/invoices-receipts.html

Amazon Textract does not support asynchronous currently for Analyze expense(check not in above documentation). We have taken note of your request. Various factors affect feature releases, and we are unable to provide specific release times at this point. In the meantime, we would be interested in hearing the details of your use case.

AWS
답변함 3년 전
0

Thanks for your reply.

Case of use
We receive an invoice as a PDF file and we need to extract as much metadata as possible, so the AnalyzeExpense functionality is perfect for us.
The main problem is that PDF files can be multipage and can also be in multiple languages.

Without async functionality we need:

  • Transform pdf into image file (each page)
  • Send to Api each one (and wait for each response)
  • Join all results

The problem with languages other than English is that most of the labels (in the fields of the order lines) are not recognized propertly, is there an option to tell the system which labels or which language to use?
I mean, can I tell the system that the label "Quantity" for me (in this operation) will be "Cantidad"?

Regards

Edited by: fcarbajal on Sep 23, 2021 3:40 AM

답변함 3년 전
0

Thank you fcarbajal@ for use case.

I have forwarded your feedback to the right team.
Stay tuned for more updates from AWS Textract!

AWS
답변함 3년 전

로그인하지 않았습니다. 로그인해야 답변을 게시할 수 있습니다.

좋은 답변은 질문에 명확하게 답하고 건설적인 피드백을 제공하며 질문자의 전문적인 성장을 장려합니다.

질문 답변하기에 대한 가이드라인

관련 콘텐츠