Gaussian Variation Field Diffusion for High-fidelity Video-to-4D Synthesis

TL;DR

This paper introduces a novel framework for high-fidelity video-to-4D synthesis that efficiently encodes 3D content and its dynamics, enabling high-quality animated 3D generation from single videos with strong generalization capabilities.

Contribution

It proposes a Direct 4DMesh-to-GS Variation Field VAE and a Gaussian Variation Field diffusion model conditioned on videos, advancing 3D animation synthesis from limited data.

Findings

Superior generation quality over existing methods

Effective generalization to in-the-wild videos

Trained on synthetic data but performs well on real videos

Abstract

In this paper, we present a novel framework for video-to-4D generation that creates high-quality dynamic 3D content from single video inputs. Direct 4D diffusion modeling is extremely challenging due to costly data construction and the high-dimensional nature of jointly representing 3D shape, appearance, and motion. We address these challenges by introducing a Direct 4DMesh-to-GS Variation Field VAE that directly encodes canonical Gaussian Splats (GS) and their temporal variations from 3D animation data without per-instance fitting, and compresses high-dimensional animations into a compact latent space. Building upon this efficient representation, we train a Gaussian Variation Field diffusion model with temporal-aware Diffusion Transformer conditioned on input videos and canonical GS. Trained on carefully-curated animatable 3D objects from the Objaverse dataset, our model demonstrates superior generation quality compared to existing methods. It also exhibits remarkable generalization to in-the-wild video inputs despite being trained exclusively on synthetic data, paving the way for generating high-quality animated 3D content. Project page: https://gvfdiffusion.github.io/.