ABC: A Big CAD Model Dataset for Geometric Deep Learning

By Sebastian Koch, Albert Matveev, Zhongshi Jiang, Francis Williams, Alexey Artemov, Evgeny Burnaev, Marc Alexa, Denis Zorin, Daniele Panozzo. 2018

678 citations recorded by the source corpus.

We introduce ABC-Dataset, a collection of one million Computer-Aided Design (CAD) models for research of geometric deep learning methods and applications. Each model is a collection of explicitly parametrized curves and surfaces, providing ground truth for differential quantities, patch segmentation, geometric feature detection, and shape reconstruction. Sampling the parametric descriptions of surfaces and curves allows generating data in different formats and resolutions, enabling fair comparisons for a wide range of geometric learning algorithms. As a use case for our dataset, we perform a large-scale benchmark for estimation of surface normals, comparing existing data driven methods and evaluating their performance against both the ground truth and traditional normal estimation methods.

Subfields: Computational Mechanics, Industrial and Manufacturing Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

1. Rundi Wu
2. Jitendra Malik
3. Ali Farhadi
4. Jordi Salvador
5. Ludwig Schmidt
6. Jiahao Li
7. Xiaodan Du
8. Haochen Wang

In Industrial and Manufacturing Engineering

1. Jitendra Malik
2. Ali Farhadi
3. Ludwig Schmidt
4. Gregory Shakhnarovich

Citations in the corpus

• Isometric Transformation Invariant and Equivariant Graph Convolutional Networks (2020, methodology): “We selected 82 CAD shapes randomly considering the first 200 shapes of the ABC dataset (Koch et al., 2019), generate first-order tetrahedral meshes with the mesher program Gmsh (Geuzaine & Remacle, 2009), and finally, performed finite element analysis (FEA) with the FEA program FrontISTR (Morita et…”
• ABO: Dataset and Benchmarks for Real-World 3D Object Understanding (2021, background): “One way around the challenging problem of getting 3D annotations for real images is to focus only on synthetic, computer-aided design (CAD) models [5, 27, 67].”
• Score Jacobian Chaining: Lifting Pretrained 2D Diffusion Models for 3D Generation (2022, background): “Many prior works [2, 58, 60] perform 3D generative modeling by training on 3D datasets [5, 23, 55, 59].”
• Objaverse: A Universe of Annotated 3D Objects (2022, methodology): “Datasets for CAD models, such as ModelNet [78] and DeepCAD [77], and ABC [35] do not include textures or materials, which limits their ability to represent objects that could plausibly be found in the real world.”
• DeepCAD: A Deep Generative Network for Computer-Aided Design Models (2021, methodology): “In particular, the ABC dataset [22] collects about 1 million CAD designs from Onshape, a webbased CAD tool and repository [3].”
• ClipGen: A Deep Generative Model for Clipart Vectorization and Synthesis (2021, methodology): “as ShapeNet [8], ModelNet [48], PartNet [49], ABC [50], and”