MVGamba: Unify 3D Content Generation as State Space Sequence Modeling

TL;DR

MVGamba introduces a lightweight, unified 3D content generation model using a state space sequence approach, improving multi-view consistency and detail in 3D reconstructions with lower computational costs.

Contribution

It proposes a novel Gaussian reconstruction model based on RNN-like State Space Models that enhances multi-view information propagation and integrates seamlessly with diffusion models.

Findings

Outperforms state-of-the-art in 3D generation tasks

Achieves high-quality 3D content with 0.1x model size

Improves multi-view consistency and detail in reconstructions

Abstract

Recent 3D large reconstruction models (LRMs) can generate high-quality 3D content in sub-seconds by integrating multi-view diffusion models with scalable multi-view reconstructors. Current works further leverage 3D Gaussian Splatting as 3D representation for improved visual quality and rendering efficiency. However, we observe that existing Gaussian reconstruction models often suffer from multi-view inconsistency and blurred textures. We attribute this to the compromise of multi-view information propagation in favor of adopting powerful yet computationally intensive architectures (e.g., Transformers). To address this issue, we introduce MVGamba, a general and lightweight Gaussian reconstruction model featuring a multi-view Gaussian reconstructor based on the RNN-like State Space Model (SSM). Our Gaussian reconstructor propagates causal context containing multi-view information for cross-view self-refinement while generating a long sequence of Gaussians for fine-detail modeling with linear complexity. With off-the-shelf multi-view diffusion models integrated, MVGamba unifies 3D generation tasks from a single image, sparse images, or text prompts. Extensive experiments demonstrate that MVGamba outperforms state-of-the-art baselines in all 3D content generation scenarios with approximately only $0.1\times$ of the model size.