OCR POC with PDFTron

[m-prodan]

Title of ticket:

Description

This is a key dependency for #4120 - we know PDFtron uses a third party OCR library, can we do the same and make it work to OCR image files?

Dependencies

Are there any dependencies?

DOD

• [ ] Analyze for Build vs Buy - PDFTron's IRIS OCR vs Custom OCR lib from us - Create samples of Samples(Handwritten, Scanned Printed etc. from CFD team)
• [ ] See what options are to export OCR'd records so that they are searchable by interest holders outside the system
• [ ] Is PDFTRON's Export TExt exposed in API to use our Custom OCR ? if not, buy is only option
• [ ] Research and Choose Library
• [ ] Create POC for OCR integration with PDFTron

[lmullane]

Check to see if you own OCR library can highlight search terms in the redaction app.

Try scanned PDFs and handwritten notes.

[nkan-aot]

OCR POC Analysis.xlsx

[nkan-aot]

I have uploaded my analysis notes above. Moving this task to Review and we will wait for the PDFTron demo before making a final decision

[abin-aot]

Need to analyze PDFTON's Backend OCR library and other Tessaract Front end options. Need to re-estimate as we start this task, moving back to Product Backlog cc: @nkan-aot , @m-prodan

[abin-aot]

New thought as discussed, with using Scanner/Hardware with any availble software to Scan directly to PDF(a must!) and #2, if possible convert those PDFs to searchable PDFs , so that we can use it on our DocReviewer App. Need to try out things on FISGARD office on machine that has access to Scanning Team's scanners and better to use their exact machine to try out options utilizing their systems resources - in other words, to do this my IDIR need access to those scanners, work stations cc: @lmullane , @m-prodan

[nkan-aot]

Here is additional analysis after meeting with PDFTron OCR comparison.xlsx python POC's have been uploaded to ms teams dev channel folder tesseract.py POC has been pushed to branch dev-NK-4238