canonical graph construction

This study addresses the limited robustness and poor engineering applicability of traditional modal shape identification methods, which heavily rely on manual expertise and struggle with variations in vehicle models, finite element meshes, and sensor layouts. To overcome these challenges, this work proposes an engineering semantics–driven graph neural network framework that constructs a geometry-agnostic canonical engineering graph representation, thereby unifying heterogeneous finite element models and experimental data while decoupling engineering knowledge from numerical discretization to enable cross-vehicle transferability. The approach incorporates a region-aware graph attention mechanism, engineering relation–guided graph pooling, and geometry-invariant regional descriptors to achieve physically interpretable 3D modal shape recognition. Validation across four vehicle datasets demonstrates high classification accuracy and strong generalization across vehicle types—even under label scarcity—with identified regions showing close alignment with established NVH engineering zones.

This work addresses the common failure of geometric approximation during mesh generation from parametric boundary representation (B-Rep) models, which often corrupts the original topological structure and leads to incorrect adjacency relationships. The paper proposes a topology-first meshing approach that, for the first time, enforces B-Rep topology as a hard constraint at the algorithmic level, rigorously preserving topological invariance throughout the discretization process. Geometric deviation is controlled solely through user-defined tolerances, thereby decoupling topological correctness from geometric approximation. The method produces robust, topologically consistent meshes without requiring post-processing and has been validated on thousands of real-world CAD models—including cases where conventional tools fail—demonstrating significantly improved reliability for downstream applications.

Existing CAD learning approaches discretize B-Rep models into triangle meshes, thereby discarding the analytical surface representations and topological information essential for consistent instance-level analysis. This work proposes STEP-Parts, a deterministic pipeline that directly extracts geometric instance partitions from native STEP B-Rep data. The method defines partitions based on intrinsic B-Rep topology, merges faces using analytical surface types and near-tangent plane continuity criteria, and transfers labels to triangulated meshes via face-to-mesh correspondence mapping. STEP-Parts ensures boundary consistency across varying triangulations and processes the DeepCAD subset of the ABC dataset—comprising approximately 180,000 models—in under six hours. The resulting labels significantly enhance performance in implicit reconstruction-segmentation tasks and point cloud networks. Code and precomputed labels are publicly released.

This work addresses the limitations of existing text-to-CAD generation methods, which often neglect assembly hierarchies and geometric constraints, leading to an excessively large search space and error accumulation. To overcome these challenges, the authors propose a hierarchical, geometry-aware graph representation that models parts and subassemblies as nodes and encodes geometric constraints as edges. The framework first predicts the structural layout and associated constraints, then leverages this information to guide the generation of CAD modeling operations and code. A novel structure-aware progressive curriculum learning strategy is introduced, employing controlled editing to construct tiered tasks and synthesize boundary cases. Additionally, the study presents the first Text-to-CAD dataset annotated with exploded views and explicit geometric constraints, along with tailored evaluation metrics. Experiments demonstrate that the proposed method significantly outperforms existing approaches in geometric fidelity and constraint satisfaction, with validation on a newly curated dataset of 12K samples.

Existing methods for generating manifold meshes typically rely on indirect representations—such as level sets or template deformations—making it difficult to directly produce high-quality, topologically unconstrained polygonal meshes with structural integrity. This paper introduces the first end-to-end differentiable framework that explicitly models half-edge structure via vertex-level continuous connectivity embeddings, enabling direct generation of discrete manifold-conforming meshes in a continuous latent space. Key contributions include: (1) the first continuous neighborhood relation learning mechanism; (2) mesh distribution fitting via stochastic optimization; and (3) topology-agnostic generation and repair capabilities. Evaluated on large-scale datasets, our method significantly improves mesh element quality, geometric fidelity, and topological diversity. It establishes the first truly end-to-end differentiable approach for manifold mesh generation and repair, bridging a critical gap between implicit representation learning and explicit, valid mesh synthesis.

Existing CAD generation methods struggle to simultaneously preserve modeling history, topological reference stability, and feature-level editability in cross-platform scenarios. This work proposes CADIR—an agent-oriented, executable intermediate representation that explicitly constructs a procedural graph encompassing operation sequences, parameter dependencies, constraints, and topological selections based on the OpenCASCADE (OCCT) geometric kernel. To enable faithful cross-platform model reconstruction, CADIR introduces a geometric signature matching mechanism. It is the first approach to support explicit procedural graph representations that allow editing across heterogeneous CAD backends. By integrating text- or image-driven procedural graph retrieval, CADIR demonstrates high-fidelity, editable reuse of complete models and substructures across FreeCAD, SolidWorks, and Fusion 360, enabling seamless subsequent modifications.

Traditional engineering design relies heavily on costly simulations, while existing data-driven approaches often overlook the parametric nature and design semantics of CAD models, limiting their integration into design workflows and interpretability. This work proposes Attribute Feature Graphs (AFGs), which, for the first time, encode native CAD features—such as extrusions and ribs—as graph nodes, with directed edges representing geometric and dependency relationships. This representation preserves design intent while enabling end-to-end learning with graph neural networks (GNNs). Evaluated on the CarHoods10K dataset, the resulting GNN surrogate model achieves prediction accuracy comparable to state-of-the-art methods and supports direct feature editing within CAD environments with real-time performance feedback. Crucially, the approach provides traceability and interpretability by linking predictions back to specific design features.

Existing topology-aware mesh generation methods couple geometry and topology within a shared latent space, often leading to vertex drift and surface discontinuities. This work proposes a decomposed flow-matching framework that first generates high-fidelity vertices based on a shared coarse voxel scaffold and subsequently produces connectivity conditioned on these vertices. The approach employs two dedicated variational autoencoders: one for sub-voxel-level vertex reconstruction and another for continuous latent embedding of discrete connectivity patterns. This design enables part-wise generation and automatic topological adaptation. Experimental results demonstrate that the proposed method outperforms state-of-the-art approaches in both geometric fidelity and connection quality.

Existing text-to-CAD generation methods struggle to maintain structural consistency and accurately realize geometric parameters in complex designs. This work proposes HierCAD, a framework that formulates CAD generation as a hierarchical reasoning process: high-level object-wise procedural reasoning is coupled with low-level part-wise topological reasoning. By integrating a joint learning mechanism for structural alignment and parameter realization—augmented with large language models, CAD construction tree decomposition, parameter perturbation, and ranking-based supervision—the approach effectively mitigates shortcut learning. Experimental results demonstrate that HierCAD outperforms state-of-the-art methods on both CAD sequence generation and reconstruction tasks, achieving significant improvements in structural fidelity and parameter accuracy.