EDIT: Here's a quick video showing how this was done: [ Ссылка ]
This is a test for the new Metahuman Animator facial mocap system from @UnrealEngine - and a couple other things. I ran a stylized head mesh through the new Mesh-to-Metahuman process (which is way better at maintaining the stylized shape compared to the old version) and animated with a video take of myself making some exaggerated faces. Note - this has zero mocap cleanup - just the straight take. The head is textured using the Midjourney generative AI - I fed a blank render of the head mesh into Midjourney with a prompt for a handpainted texture, then projection mapped it onto the mesh in Blender, and then handpainted the parts that were messy (the sides and back, inside the mouth etc).
![](https://s2.save4k.ru/pic/D8XK4Ugqzm8/maxresdefault.jpg)