Nice job!
I played around with the UI, obviously the quality of images could use improvement, but you've already mentioned that it's a low quality to keep the UI quick. Although this may eventually replace most 3D asset creation, I wonder if it can be used by 3D artists as a starting point to their modeling work.
I'm psyched to try this! Was looking for something like this a long time ago when working on a project.
One feedback on the name - Is this primarily for Home decor category ? if not maybe the name Homely is a bit limiting. I see the UI demo only has sofas, are you going to be adding other categories of products?
Thanks, do let me know any feedback when you try it.
Regarding the name, yeah we'll probably need to pick something more generic. Right now though it's only home decor categories like sofa, chairs but the idea is to expand to broader categories.
Thanks for trying it out !
Ah, yes. It doesn't work great with sectionals yet. The model does an Ok job on the kind of sofa styles it's seen during training but still needs to generalize better. Also the UI demo is a quick sub second inference. But the quality gets progressively better with iterative refinement via the API which isn't happening right now.
Looks awesome. Do you mind if I ask how this works behind the hood? I assume some kind of computer vision ML algorithms are involved, probably GANs? But how are you going from the 2D image to the 3D realm?
7 comments
[ 4.0 ms ] story [ 21.3 ms ] threadTotal fail.
[1] https://www.google.com/search?q=cor+trio+sofa