Search papers, labs, and topics across Lattice.
HierCAD introduces a hierarchical framework for text-to-CAD design that enhances structural reasoning and parameter prediction by reformulating CAD generation into progressive reasoning stages. The method employs a unified Structure Alignment and Parameter Grounding (SAPG) strategy, which aligns topology reasoning with parametric CAD spans and mitigates shortcut learning through structured perturbations. Experimental results show that HierCAD significantly outperforms existing state-of-the-art approaches in both CAD sequence generation and the evaluation of reconstructed CAD models.
HierCAD achieves unprecedented fidelity in CAD generation by aligning structural reasoning with geometric parameters, setting a new benchmark for text-to-CAD systems.
Recent text-to-CAD approaches have shown promising results by leveraging large language models, but they often struggle with maintaining structural consistency in complex designs and accurately grounding geometric parameters. To address these issues, we propose HierCAD, a hierarchical text-to-CAD framework that improves both structural reasoning and parameter prediction. HierCAD reformulates CAD generation as progressive reasoning by decomposing CAD construction trees into object-level procedural reasoning and part-level topology reasoning trajectories. To further improve generation fidelity, we introduce a unified Structure Alignment and Parameter Grounding (SAPG) learning strategy. Structure alignment aligns topology reasoning trajectories with their corresponding parametric CAD spans, while parameter grounding mitigates shortcut learning through structure-preserving parameter perturbations and ranking-based supervision. Experiments demonstrate that HierCAD outperforms prior state-of-the-art methods on both CAD sequence generation and reconstructed CAD model evaluation. Our code is available at https://github.com/Collab-Gen/HierCAD.
