Let 2D Diffusion Model Know 3D-Consistency for Robust Text-to-3D Generation

TL;DR

This paper introduces 3DFuse, a framework that enhances pretrained 2D diffusion models with 3D awareness, improving the robustness and consistency of text-to-3D scene generation using score distillation.

Contribution

The paper proposes a novel method to incorporate 3D awareness into 2D diffusion models, enabling more stable and 3D-consistent text-to-3D generation.

Findings

Improved 3D consistency in generated scenes.

Enhanced robustness against errors in coarse 3D structures.

Surpassed prior methods in quality of 3D scene reconstruction.

Abstract

Text-to-3D generation has shown rapid progress in recent days with the advent of score distillation, a methodology of using pretrained text-to-2D diffusion models to optimize neural radiance field (NeRF) in the zero-shot setting. However, the lack of 3D awareness in the 2D diffusion models destabilizes score distillation-based methods from reconstructing a plausible 3D scene. To address this issue, we propose 3DFuse, a novel framework that incorporates 3D awareness into pretrained 2D diffusion models, enhancing the robustness and 3D consistency of score distillation-based methods. We realize this by first constructing a coarse 3D structure of a given text prompt and then utilizing projected, view-specific depth map as a condition for the diffusion model. Additionally, we introduce a training strategy that enables the 2D diffusion model learns to handle the errors and sparsity within the coarse 3D structure for robust generation, as well as a method for ensuring semantic consistency throughout all viewpoints of the scene. Our framework surpasses the limitations of prior arts, and has significant implications for 3D consistent generation of 2D diffusion models.