Mesh-RFT can generate product-ready meshes when conditioned on point cloud derived from dense meshes generated by Hunyuan3D.
Existing pretrained models for 3D mesh generation often suffer from data biases and produce low-quality results, while global reinforcement learning (RL) methods rely on object-level rewards that struggle to capture local structure details. To address these challenges, we present Mesh-RFT, a novel fine-grained reinforcement fine-tuning framework that employs Masked Direct Preference Optimization (M-DPO) to enable localized refinement via quality-aware face masking. To facilitate efficient quality evaluation, we introduce an objective topology-aware scoring system to evaluate geometric integrity and topological regularity at both object and face levels through two metrics: Boundary Edge Ratio (BER) and Topology Score (TS). By integrating these metrics into a fine-grained RL strategy, Mesh-RFT becomes the first method to optimize mesh quality at the granularity of individual faces, resolving localized errors while preserving global coherence. Experiment results show that our M-DPO approach reduces Hausdorff Distance (HD) by 24.6% and improves Topology Score (TS) by 3.8% over pre-trained models, while outperforming global DPO methods with a 17.4% HD reduction and 4.9% TS gain. These results demonstrate Mesh-RFT's ability to improve geometric integrity and topological regularity, achieving new state-of-the-art performance in production-ready mesh generation.
The proposed Mesh-RFT Framework. The pipeline comprises three stages: 1) Mesh Generation Pre-training using an Hourglass AutoRegressive Transformer and a Shape Encoder; 2) Preference Dataset Construction where a pretrained model generates candidate meshes, and a topology-aware score system establishes preference pairs; and 3) Mesh Generation Post-training which employs Mask DPO with reference and policy networks for subsequent refinement.
During training, meshes are arranged in a bottom-up manner, resulting in generated meshes being produced in a sequential bottom-to-top order.
Generated Mesh Conditioned on point cloud from Artist Mesh
Generated Mesh Conditioned on point cloud from Dense Mesh
@misc{liu2025meshrftenhancingmeshgeneration,
title={Mesh-RFT: Enhancing Mesh Generation via Fine-grained Reinforcement Fine-Tuning},
author={Jian Liu and Jing Xu and Song Guo and Jing Li and Jingfeng Guo and Jiaao Yu and Haohan Weng and Biwen Lei and Xianghui Yang and Zhuo Chen and Fangqi Zhu and Tao Han and Chunchao Guo},
year={2025},
eprint={2505.16761},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2505.16761},
}