[hyperbrainer]

I wonder whether CV models could help speed this up. There is the obvious challenge of pieces not fitting together exactly, being broken or otherwise defaced etc. but then humans are struggling with the same thing, just that computers have far better memories.