GeoFusion-CAD: Structure-Aware Diffusion with Geometric State Space for Parametric 3D Design

2026-03-23

AI Summary

Existing methods struggle to generate long, parameterized CAD sequences with complex geometric and topological dependencies, while Transformer-based approaches are hindered by quadratic attention costs and limited context length. This work proposes the first end-to-end diffusion framework based on state space models, encoding CAD programs as hierarchical tree structures and modeling them within a joint geometric-topological state space. We introduce a lightweight C-Mamba module to efficiently capture long-range dependencies and develop a structure-aware diffusion mechanism. To support comprehensive evaluation, we release DeepCAD-240, a new benchmark dataset featuring sequences of up to 240 command steps. Experiments demonstrate that our method significantly outperforms existing Transformer models in both short and long sequence generation, achieving state-of-the-art performance in geometric fidelity and topological consistency.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Parametric Computer-Aided Design (CAD) is fundamental to modern 3D modeling, yet existing methods struggle to generate long command sequences, especially under complex geometric and topological dependencies. Transformer-based architectures dominate CAD sequence generation due to their strong dependency modeling, but their quadratic attention cost and limited context windowing hinder scalability to long programs. We propose GeoFusion-CAD, an end-to-end diffusion framework for scalable and structure-aware generation. Our proposal encodes CAD programs as hierarchical trees, jointly capturing geometry and topology within a state-space diffusion process. Specifically, a lightweight C-Mamba block models long-range structural dependencies through selective state transitions, enabling coherent generation across extended command sequences. To support long-sequence evaluation, we introduce DeepCAD-240, an extended benchmark that increases the sequence length ranging from 40 to 240 while preserving sketch-extrusion semantics from the ABC dataset. Extensive experiments demonstrate that GeoFusion-CAD achieves superior performance on both short and long command ranges, maintaining high geometric fidelity and topological consistency where Transformer-based models degrade. Our approach sets new state-of-the-art scores for long-sequence parametric CAD generation, establishing a scalable foundation for next-generation CAD modeling systems. Code and datasets are available at GitHub.

Problem

• parametric CAD
• long command sequences
• geometric dependencies
• topological dependencies
• scalability

Innovation

• diffusion model
• state space model
• parametric CAD
• long-sequence generation
• C-Mamba