Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

HistCAD: Geometrically Constrained Parametric History-based CAD Dataset

Xintong Dong, Chuanyang Li, Chuqi Han, Peng Zheng, Jiaxin Jing, Yanzhi Song, Zhouwang Yang

TL;DR

HistCAD is presented, a large-scale dataset featuring constraint-aware modeling sequences that compactly represent procedural operations while ensuring compatibility with native CAD software, encompassing five aligned modalities: modeling sequences, multi-view renderings, STEP-format B-reps, native parametric files, and textual annotations.

Abstract

Parametric computer-aided design (CAD) modeling is fundamental to industrial design, but existing datasets often lack explicit geometric constraints and fine-grained functional semantics, limiting editable, constraint-compliant generation. We present HistCAD, a large-scale dataset featuring constraint-aware modeling sequences that compactly represent procedural operations while ensuring compatibility with native CAD software, encompassing five aligned modalities: modeling sequences, multi-view renderings, STEP-format B-reps, native parametric files, and textual annotations. We develop AM\(_\text{HistCAD}\), an annotation module that extracts geometric and spatial features from modeling sequences and uses a large language model to generate complementary annotations of the modeling process, geometric structure, and functional type. Extensive evaluations demonstrate that HistCAD's explicit constraints, flattened sequence format, and multi-type annotations improve robustness, parametric editability, and accuracy in text-driven CAD generation, while industrial parts included in HistCAD further support complex real-world design scenarios. HistCAD thus provides a unified benchmark for advancing editable, constraint-aware, and semantically enriched generative CAD modeling.

HistCAD: Geometrically Constrained Parametric History-based CAD Dataset

TL;DR

HistCAD is presented, a large-scale dataset featuring constraint-aware modeling sequences that compactly represent procedural operations while ensuring compatibility with native CAD software, encompassing five aligned modalities: modeling sequences, multi-view renderings, STEP-format B-reps, native parametric files, and textual annotations.

Abstract

Parametric computer-aided design (CAD) modeling is fundamental to industrial design, but existing datasets often lack explicit geometric constraints and fine-grained functional semantics, limiting editable, constraint-compliant generation. We present HistCAD, a large-scale dataset featuring constraint-aware modeling sequences that compactly represent procedural operations while ensuring compatibility with native CAD software, encompassing five aligned modalities: modeling sequences, multi-view renderings, STEP-format B-reps, native parametric files, and textual annotations. We develop AM, an annotation module that extracts geometric and spatial features from modeling sequences and uses a large language model to generate complementary annotations of the modeling process, geometric structure, and functional type. Extensive evaluations demonstrate that HistCAD's explicit constraints, flattened sequence format, and multi-type annotations improve robustness, parametric editability, and accuracy in text-driven CAD generation, while industrial parts included in HistCAD further support complex real-world design scenarios. HistCAD thus provides a unified benchmark for advancing editable, constraint-aware, and semantically enriched generative CAD modeling.
Paper Structure (27 sections, 1 equation, 8 figures, 6 tables, 2 algorithms)

This paper contains 27 sections, 1 equation, 8 figures, 6 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. CAD Model Representations
  4. Generative Modeling for CAD
  5. 3D Content Annotation
  6. HistCAD Dataset
  7. Creation of the HistCAD Dataset
  8. HistCAD-Academic
  9. HistCAD-Industrial
  10. Modeling Sequence Representation
  11. Sketch
  12. Extrusion
  13. Boolean Operations
  14. Execution in CAD Environments
  15. Statistical Analysis and Performance Evaluation
  16. ...and 12 more sections

Figures (8)

  • Figure 1: A random sampling of designs from HistCAD.
  • Figure 2: Token count distributions across datasets (capped at 4096 tokens).
  • Figure 3: Qualitative comparison of reconstructed shapes. Each row shows reconstructions from HistCAD, Text2CAD khan2024text2cad, and DeepCAD wu_2021_iccv, with the ground-truth ABC model shown on the right.
  • Figure 4: Editing results with and without geometric constraints. For each model, we compare an edited version with all original constraints removed against one preserving HistCAD’s constraints. Coincidence constraints are automatically added in both cases to ensure closed loops.
  • Figure 5: Qualitative comparison of CAD models generated by Text2CAD$_T$ and HistCAD$_T$ using Text2CAD’s abstract and beginner-level annotations.
  • ...and 3 more figures