Tencent/Hunyuan3D-1

Any Future plans to have multiview image sequences to improve the generated model?

Opened this issue · 0 comments

I should elaborate on this being in use cases where there is either:

  • a 360 degree view image sequence or video of a static object with semi sparse frame sampling
  • a 360 degree view image sequence or video of a "slightly" animated object with semi sparse frame sampling
  • or a non-360 degree view sequence or video that might have slight image variance of a "slightly" animated object