2D Diffusion Models for 3D Generation: How They're Related to Wonder3D

Table of Links

Abstract and 1 Introduction

2. Related Works

2.1. 2D Diffusion Models for 3D Generation

2.2. 3D Generative Models and 2.3. Multi-view Diffusion Models

3. Problem Formulation

3.1. Diffusion Models

3.2. The Distribution of 3D Assets

4. Method and 4.1. Consistent Multi-view Generation

4.2. Cross-Domain Diffusion

4.3. Textured Mesh Extraction

5. Experiments

5.1. Implementation Details

5.2. Baselines

5.3. Evaluation Protocol

5.4. Single View Reconstruction

5.5. Novel View Synthesis and 5.6. Discussions

6. Conclusions and Future Works, Acknowledgements and References

2. Related Works

2.1. 2D Diffusion Models for 3D Generation

Recent compelling successes in 2D diffusion models [8, 22, 47] and large vision language models (e.g., CLIP model [45]) provide new possibilities for generating 3D assets using the strong priors of 2D diffusion models. Pioneering works DreamFusion [43] and SJC [59] propose to distill a 2D text-to-image generation model to generate 3D shapes from texts, and many follow-up works follow such per-shape optimization scheme.

For the task of textto-3D [2, 5, 6, 23, 29, 48, 49, 57, 63, 65, 69, 77] or imageto-3D synthesis [38, 44, 46, 50, 54, 67], these methods typically optimize a 3D representation (i.e., NeRF, mesh, or SDF), and then leverage neural rendering to generate 2D images from various viewpoints. The images are then fed into the 2D diffusion models or CLIP model for calculating SDS [43] losses, which can guide the 3D shape optimization.

However, most of these methods always suffer from low efficiency and multi-face problem, where a per-shape optimization consumes tens of minutes and the optimized geometry tends to produce multiple faces due to the lack of explicit 3D supervision. A recent work one-2-3-45 [15] proposes to leverage a generalizable neural reconstruction method SparseNeuS [36] to directly produce 3D geometry from the generated images from zero123 [31]. Although the method achieves high efficiency, its results are of low-quality and lack geometric details.