I am using AI Builder - Form processing option for invoice pdf data extraction. In the document, we have product lines in the table format. If document has more than three product lines, the table data is capturing. But if one or two product lines are there, the table data is not capturing. I have uploaded 47 documents of same format pdf with 4,3,2 and 1 product lines. How to capture data of invoice pdf with 1 or 2 product lines.
If the invoices don't contain confidential information, would you be able to share them with me so the AI Builder team can analyze why table data is not being captured correctly? You can share them via a private message with me if you prefer.
I have modified the invoice with dummy data and shared with you as zip file. I have used these invoices for form processing. Kindly find the attachment.
Thanks @Saratha30_M, this is really helpful.
Just to make sure that I understood correctly, what you would like is to have the area in the purple rectangle be captured as a table?
Exactly Joe. Once extracting the invoice, prediction output of tables is empty as below.
Kindly find the attached file of pdf documents with 2 product lines. I have created with dummy data and checked with model form processing. This is analysing the table data. But invoice date and number is not analysing. In this, two product lines with original pdf document is not analysing table data. I couldn't share the pdf.
Thank you for providing the samples.
Yes, with these new samples the table is correctly detected.
For date and invoice # to be detected, the values will need to change in some of the documents. In the samples you provided they are always the same: "June 10, 2019" and "INV/19-20/10001"
For the original data set, make sure you have invoices with multiple product lines to have it detected as a table.
Thanks for your suggestion. I will change the date, invoice and check the same. Is one product line invoice capture the table data. Could you please suggest if any wrong in the pdf document.
Continue your learning in our online communities.
Next Wednesday, September 18th at 8am PDT
Features releasing from October 2019 through March 2020
Coming to a city near you
Fill out a quick form to claim your user group badge now!
Connect, share, and learn with your peers year-round
Register by September 5 to save $200