AI UV unwrapping

I saw this idea floating around somewhere on the forum and decided it deserves it’s own topic.

I’m just wondering with constant expansion of AI R&D why we still don’t have AI unwrappers?
Is UV field so unattractive, or are there any hidden limitations that prevent it?

To me it looks like a match made in heaven. UV unwrap is an unsolvable problem, that only requires good-enough solution, and there are tons of training data and examples how good UV should look. There are very specific measurable criteria of UV quality (texel density, UV stretch). I mean, come on.

I HATE unwrapping and would gladly see it automated.

1 Like

Perhaps is like scientists have become too obsessed about hitting trendy subjects to become rich with topics such as cool image generation etc. The boring and labor intensive work that goes on the backend, no one gives a dime.

Not to mention about, weight painting, skinning, armature posing (perhaps Cascadeur? is a solution), vertex-edge smoothing, surface generation… Too many topics need to be tackled at once.

2 Likes