[emj]

Since most photos of cuneiform will be professionally taken it's going to be a bit easier. There are terabytes of images, and apparently lots of tagged data to be used [1] so not an impossible project.

Just now I learned that standard is to take photos with lighting from the top left corner to be able to read these [2], that said reading physical clay tablets is a bit harder than just reading a paper. I have often been surprised that 3D objects are so hard to interpret correctly from just photos.

[1] https://github.com/oi-deepscribe/deepscribe [2] https://youtu.be/dwGmyy2Aabg?t=690