From Prompts to Printable Models: Support-Effective 3D Generation via Offset Direct Preference Optimization

2025-11-20

AI Summary

To address the high material cost and labor-intensive post-processing caused by support structures in 3D printing of overhanging geometries, this paper proposes SEG—the first end-to-end, support-aware text-to-printable 3D model generation framework. Our core innovation introduces Offset-augmented Direct Preference Optimization (ODPO) into 3D generative modeling for the first time, enabling joint training of a physics-based support structure simulator and the generative model. This co-training implicitly optimizes generated geometry during inference to minimize support requirements—eliminating the need for post-hoc support removal or manual editing. Evaluated on Thingi10k-Val and GPT-3DP-Val, SEG outperforms TRELLIS, DPO, and DRO baselines, achieving an average 32.7% reduction in support volume and a 19.4% improvement in print success rate, while preserving high prompt fidelity.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Sampling/Simulation-based Search
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

The transition from digital 3D models to physical objects via 3D printing often requires support structures to prevent overhanging features from collapsing during the fabrication process. While current slicing technologies offer advanced support strategies, they focus on post-processing optimizations rather than addressing the underlying need for support-efficient design during the model generation phase. This paper introduces SEG ( extit{underline{S}upport-underline{E}ffective underline{G}eneration}), a novel framework that integrates Direct Preference Optimization with an Offset (ODPO) into the 3D generation pipeline to directly optimize models for minimal support material usage. By incorporating support structure simulation into the training process, SEG encourages the generation of geometries that inherently require fewer supports, thus reducing material waste and production time. We demonstrate SEG's effectiveness through extensive experiments on two benchmark datasets, Thingi10k-Val and GPT-3DP-Val, showing that SEG significantly outperforms baseline models such as TRELLIS, DPO, and DRO in terms of support volume reduction and printability. Qualitative results further reveal that SEG maintains high fidelity to input prompts while minimizing the need for support structures. Our findings highlight the potential of SEG to transform 3D printing by directly optimizing models during the generative process, paving the way for more sustainable and efficient digital fabrication practices.

Problem

• Optimizing 3D models for minimal support material usage
• Reducing material waste and production time in 3D printing
• Generating printable geometries with high fidelity to prompts

Innovation

• Integrates Offset Direct Preference Optimization into 3D generation
• Simulates support structures during training to reduce waste
• Optimizes models for minimal support material usage directly