V3D: Video Diffusion Models are Effective 3D Generators

TL;DR

V3D introduces a novel 3D generation method leveraging pre-trained video diffusion models, achieving high-quality, multi-view consistent 3D objects and scenes with fast reconstruction times.

Contribution

The paper presents V3D, a new approach that adapts video diffusion models for effective 3D object and scene generation with geometric consistency and rapid reconstruction.

Findings

State-of-the-art multi-view 3D generation quality

High-fidelity 3D meshes and Gaussians generated within 3 minutes

Effective scene-level novel view synthesis from sparse inputs

Abstract

Automatic 3D generation has recently attracted widespread attention. Recent methods have greatly accelerated the generation speed, but usually produce less-detailed objects due to limited model capacity or 3D data. Motivated by recent advancements in video diffusion models, we introduce V3D, which leverages the world simulation capacity of pre-trained video diffusion models to facilitate 3D generation. To fully unleash the potential of video diffusion to perceive the 3D world, we further introduce geometrical consistency prior and extend the video diffusion model to a multi-view consistent 3D generator. Benefiting from this, the state-of-the-art video diffusion model could be fine-tuned to generate 360degree orbit frames surrounding an object given a single image. With our tailored reconstruction pipelines, we can generate high-quality meshes or 3D Gaussians within 3 minutes. Furthermore, our method can be extended to scene-level novel view synthesis, achieving precise control over the camera path with sparse input views. Extensive experiments demonstrate the superior performance of the proposed approach, especially in terms of generation quality and multi-view consistency. Our code is available at https://github.com/heheyas/V3D