Imagine if you could take this picture, as well as a backshot or back photo and have this thing analyze both here and construct them into one mesh?
@ilova2597 Жыл бұрын
Can you make tutorial ?
@carlmarcs36479 ай бұрын
kzbin.info/www/bejne/qnOvnGeopKmHebM here's a pretty good tutorial on it
@wasdwasd6098 ай бұрын
AI image -> Control Mask of said image -> 100+ subdivided plane -> Add/Apply mask -> Shade Smooth -> Add AI texture = Complete. Then you can apply mirrors or whatever to make it usable if possible.
@rajendrameena15010 ай бұрын
Still not as deep as it should be.
@randaldong10 ай бұрын
Yeah, agree. The recent Wonder3D is way better :)
@jasondavis17408 ай бұрын
I've done similar experiments running 2 displacements one with the depth image and one with a normal map and it seems to help a bit with the depth issues. Also if you remove the background with an alpha channel there's a great way using geometry nodes to get rid of the geometry easily. Alternatively you could make an SVG of the silhouette turn it into a curve and knife project onto the original model and remove everything else :)