HexaGen3D: StableDiffusion is just one step away from Fast and Diverse Text-to-3D Generation
This addresses the problem of data scarcity and slow generation for 3D content creation, offering a fast and diverse solution for applications in gaming, VR, and design, though it builds incrementally on existing 2D diffusion methods.
The paper tackles the challenge of efficient text-to-3D generation by proposing HexaGen3D, which fine-tunes a pretrained 2D diffusion model to predict orthographic projections and a latent triplane, enabling high-quality and diverse 3D mesh generation from text prompts in 7 seconds without per-sample optimization.
Despite the latest remarkable advances in generative modeling, efficient generation of high-quality 3D assets from textual prompts remains a difficult task. A key challenge lies in data scarcity: the most extensive 3D datasets encompass merely millions of assets, while their 2D counterparts contain billions of text-image pairs. To address this, we propose a novel approach which harnesses the power of large, pretrained 2D diffusion models. More specifically, our approach, HexaGen3D, fine-tunes a pretrained text-to-image model to jointly predict 6 orthographic projections and the corresponding latent triplane. We then decode these latents to generate a textured mesh. HexaGen3D does not require per-sample optimization, and can infer high-quality and diverse objects from textual prompts in 7 seconds, offering significantly better quality-to-latency trade-offs when comparing to existing approaches. Furthermore, HexaGen3D demonstrates strong generalization to new objects or compositions.