spatialcomputing.info

I am learning ARKit and would like to work through a tutorial on giving a user virtual glasses (could be earrings, I guess, but I have glasses .obj files to use). I know this involves face detection, but I’m not sure how to then anchor it to the user’s movements. If I’m getting really advanced I would want the user to be able to push them up or down on their nose, but that might be asking too much.

If you have the 3D object for the glasses, do you just need to use ARKit/ARCore to scan someone’s face and then that’s it?

Has anyone heard of or tried Ditto? They supposedly make their own tech for understanding facial geometry to do glasses try ons, but I don’t understand how existing solutions aren’t good enough. I also tried their solution on the Bose website and it didn’t seem very good.

I’ve played with many but found that after a few days the novelty wears off. Which ones might I be missing?