SketchGraphs: A Large-Scale Dataset for Modeling Relational Geometry in Computer-Aided Design
TLDR
Introduces SketchGraphs, a large-scale dataset of 15 million parametric CAD sketches with geometric constraint graphs for machine learning.
Reasoning
The paper's strength is its large-scale, real-world dataset and open-source pipeline, enabling generative and conditional modeling benchmarks. However, it focuses only on 2D sketches, not full 3D CAD models, and lacks broader task coverage like reconstruction or text-to-CAD.
Read-first score
Read-first score 45.6, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 58.
Field roles
Rank sensitivity
Stability: volatile; rank range: 80.
Keyword Scores
Deep Analysis
Innovations
- Introduction of SketchGraphs, a large-scale dataset of 15 million parametric CAD sketches extracted from real-world models.
- Representation of each sketch as a geometric constraint graph, where nodes are geometric primitives and edges are designer-imposed constraints.
- Open-source data processing pipeline for extracting and structuring the sketches.
Methodology
The authors collected 15 million 2D sketches from real-world parametric CAD models and represented each as a geometric constraint graph. They then defined and established benchmarks for two tasks: generative modeling of sketches and conditional generation of constraints given unconstrained geometry.
Key Results
Benchmarks were established for generative modeling of sketches and for conditional generation of likely constraints from unconstrained geometry, demonstrating the dataset's utility for these tasks.