About the object normalization in v1.2

SUDO-AI-3D / zero123plus

Code repository for Zero123++: a Single Image to Consistent Multi-view Diffusion Base Model.

Apache License 2.0

1.56k stars 108 forks source link

About the object normalization in v1.2 #73

Closed bluestyle97 closed 2 months ago

bluestyle97 commented 2 months ago

Hi, I notice that the v1.2 model "always outputs a set of views assuming a normalized object size instead of changing w.r.t. the input". Could you please give more explanation on how you normalize the object when rendering the data? Did you set a fixed camera distance for the 6 views? If yes, what's the value?

Dustinpro commented 2 months ago

It is normalized to fit into a sphere with radius=.5. The fixed camera distance is 0.5 / np.tan(np.radians(30/2)).

bluestyle97 commented 2 months ago

@Dustinpro Thanks for your reply!