Controllable Text-to-3D Generation via Surface-Aligned Gaussian Splatting

Arxiv 2024
Zhiqi Li 1,2 ,   Yiming Chen 2,3 ,   Lingzhe Zhao 2 ,   Peidong Liu 2

1 Zhejiang University


2 Westlake University


3 Tongji University


While text-to-3D and image-to-3D generation tasks have received considerable attention, one important but under-explored field between them is controllable text-to-3D generation, which we mainly focus on in this work. To address this task, 1) we introduce ControlNet (MVControl), a novel neural network architecture designed to enhance existing pre-trained multi-view diffusion models by integrating additional input conditions, such as edge, depth, normal, and scribble maps. Our innovation lies in the introduction of a conditioning module that controls the base diffusion model using both local and global embeddings, which are computed from the input condition images and camera poses. Once trained, MVControl is able to offer 3D diffusion guidance for optimization-based 3D generation. And, 2) we propose an efficient multi-stage 3D generation pipeline that leverages the benefits of recent large reconstruction models and score distillation algorithm. Building upon our MVControl architecture, we employ a unique hybrid diffusion guidance method to direct the optimization process. In pursuit of efficiency, we adopt 3D Gaussians as our representation instead of the commonly used implicit representations. We also pioneer the use of SuGaR, a hybrid representation that binds Gaussians to mesh triangle faces. This approach alleviates the issue of poor geometry in 3D Gaussians and enables the direct sculpting of fine-grained geometry on the mesh. Extensive experiments demonstrate that our method achieves robust generalization and enables the controllable generation of high-quality 3D content.

Method Overview

Overview of proposed method. The multi-stage pipeline can efficiently generate high-quality textured meshes starting from a set of coarse Gaussians generated by LGM, with the input being the multi-view images generated by our MVControl. In the second stage, we employ a 2D & 3D hybrid diffusion prior for Gaussian optimization. Finally, in the third stage, we calculate the VSD loss to refine the SuGaR representation.

Text-to-3D Conditioned on Canny Edges

Text-to-3D Conditioned on Normal Maps

Text-to-3D Conditioned on Depth Maps

Text-to-3D Conditioned on User Scribbles

Diverse Controllable Multi-view Image Generation

Diverse generated multi-view images of MVControl. Our MVControl can generate diverse multi-view images with same conditions.


        title={Controllable Text-to-3D Generation via Surface-Aligned Gaussian Splatting}, 
        author={Zhiqi Li and Yiming Chen and Lingzhe Zhao and Peidong Liu},