CraftsMan: High-fidelity Mesh Generation with 3D Native Generation and Interactive Geometry Refiner

Weiyu Li,Jiarui Liu,Rui Chen,Yixun Liang,Xuelin Chen,Ping Tan,Xiaoxiao Long

2024-05-24

Abstract:We present a novel generative 3D modeling system, coined CraftsMan, which can generate high-fidelity 3D geometries with highly varied shapes, regular mesh topologies, and detailed surfaces, and, notably, allows for refining the geometry in an interactive manner. Despite the significant advancements in 3D generation, existing methods still struggle with lengthy optimization processes, irregular mesh topologies, noisy surfaces, and difficulties in accommodating user edits, consequently impeding their widespread adoption and implementation in 3D modeling software. Our work is inspired by the craftsman, who usually roughs out the holistic figure of the work first and elaborates the surface details subsequently. Specifically, we employ a 3D native diffusion model, which operates on latent space learned from latent set-based 3D representations, to generate coarse geometries with regular mesh topology in seconds. In particular, this process takes as input a text prompt or a reference image and leverages a powerful multi-view (MV) diffusion model to generate multiple views of the coarse geometry, which are fed into our MV-conditioned 3D diffusion model for generating the 3D geometry, significantly improving robustness and generalizability. Following that, a normal-based geometry refiner is used to significantly enhance the surface details. This refinement can be performed automatically, or interactively with user-supplied edits. Extensive experiments demonstrate that our method achieves high efficacy in producing superior-quality 3D assets compared to existing methods. HomePage:

Graphics,Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

This paper introduces a novel 3D modeling system called CraftsMan, aiming to address the problems of high-quality, diverse shapes, regular mesh topology, and detailed surface 3D geometry generation, while allowing users to refine the geometry interactively. Existing 3D generation methods suffer from time-consuming optimization processes, irregular meshes, surface noise, and difficulties in editing, limiting their widespread use in 3D modeling software. Inspired by the workflow of craftsmen, CraftsMan first rapidly generates rough shapes and then refines surface details.

The CraftsMan system consists of two parts: a 3D-native diffusion model, which generates rough geometries with regular topology based on intermediate generated multi-view images within a few seconds, and a geometry refiner, which significantly enhances surface details through automatic or interactive editing. The 3D-native diffusion model utilizes textual prompts or reference images to generate multi-view images, which are then used for 3D geometry generation, improving robustness and generalization. The geometry refiner includes a normal-based tool for automatic or user-guided refinement editing.

Compared to existing methods, CraftsMan can generate highly realistic complex 3D shapes within 30 seconds and supports user interaction editing, thus improving the quality of the generated 3D assets. Experiments demonstrate that CraftsMan outperforms existing techniques in generating high-quality 3D assets.