MVDiff: Scalable and Flexible Multi-View Diffusion for 3D Object Reconstruction from Single-View

arXiv:2405.03894v1 Announce Type: cross
Abstract: Generating consistent multiple views for 3D reconstruction tasks is still a challenge to existing image-to-3D diffusion models. Generally, incorporating 3D representations into diffusion model decrease the model's speed as well as generalizability and quality. This paper proposes a general framework to generate consistent multi-view images from single image or leveraging scene representation transformer and view-conditioned diffusion model. In the model, we introduce epipolar geometry constraints and multi-view attention to enforce 3D consistency. From as few …

3d object 3d reconstruction abstract arxiv challenge consistent cs.cv cs.lg diffusion diffusion model diffusion models framework general generate image multiple object paper quality scalable speed tasks type view