Spaces:
Running
on
A10G
Apply for community grant: Academic project
ZoeDepth is a generic single-image metric depth estimation model with state-of-the-art generalization capabilities.
This demo would benefit greatly from GPU hardware for better interactivity and user experience.
Hi @shariqfarooq , we have assigned a gpu to this space. Note that GPU Grants are provided temporarily and might be removed after some time if the usage is very low.
To learn more about GPUs in Spaces, please check out https://huggingface.co/docs/hub/spaces-gpus
ZoeDepth is a generic single-image metric depth estimation model with state-of-the-art generalization capabilities.
This demo would benefit greatly from GPU hardware for better interactivity and user experience.
Hi,@shariqfarooq, thanks for your open-source codes! Marvelous work! To be honest, I have tried hundreds of demo images using App, very impressive!!!
I am curious about how to rebuild the 3D mesh grid if we only have the single-view image with depth? Do we need to re-estimate the parameters of the camera?
Hi
@MagicBag
,
Thanks for your interest in our work. And thank you for your kind words.
We assume some typical camera parameters and it works fine for most cases (of course it affects the metric accuracy of 3D models). Check out https://huggingface.co/spaces/shariqfarooq/ZoeDepth/blob/962e87ab4eb0e419610ddf512841dec2558dcf7c/geometry.py#L3
Cool, it is interesting to see that a typical camera parameter should work fine. Thanks!
Hi @shariqfarooq , thank you for your amazing work. I am intrigued whether the 2D image to 3D reconstruction would work for multiple images?