OCR for blacklist images

[regoldman]

No description provided.

[vsoch]

Note that I started a project that aims to do this - https://github.com/pydicom/dicom-scraper

This would be a prime project for a graduate student or similar - it's well scoped machine learning, with plenty of data for development. @regoldman do you have a pool (or can announce at some meeting the need) for a graduate student that might be interested? I can offer to container-ize / tool-ize a suitable algorithm that the graduate student develops. The one I started might also be tweaked tested, my general feeling is that the kind of OCR we need isn't terribly "unsolved" in the scope of problems.

[danielsnider]

Hi @vsoch I am working on this using computer vision ;-). I also have machine learning students that might join in. I'll give you an update in a week and then lets chat.

[vsoch]

Awesome! Re-opening, can't wait to chat!

[vsoch]

This is out of scope for deid (and no discussion has happened for about a year) so closing issue.