fraunhoferhhi / neural-deferred-shading

Multi-View Mesh Reconstruction with Neural Deferred Shading (CVPR 2022)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Number of Multi View Images Required

AmanKishore opened this issue · comments

Hi,

This is great work!

How many multi-view images are required to generate an accurate mesh?

Have you tried using Stable Diffusion to create those multi-views?

Thanks!

Hi,

the number of required images depends a couple of factors and I guess there is no general rule. 360° coverage of the object would help because we strongly rely on mask supervision.

Regarding Stable Diffusion: we have so far verified our method only on real image data.

Best,
Markus