C33D blends a 3D model with an object category by generating a fused front view, then using texture and shape multi-view diffusion plus adaptive inversion to reconstruct a novel, consistent 3D model.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Category-Aware 3D Object Composition with Disentangled Texture and Shape Multi-view Diffusion
C33D blends a 3D model with an object category by generating a fused front view, then using texture and shape multi-view diffusion plus adaptive inversion to reconstruct a novel, consistent 3D model.