A Synthetic 3D Gear Dataset for Manufacturing Quality Inspection (MFGNet-Gear)
Quick Answer
MFGNet-Gear is a synthetic 3D dataset for quality inspection in manufacturing, featuring 24,000 paired polygon meshes and point clouds across 12 gear designs and 4 quality classes.
Quick Take
This dataset addresses the challenges of class imbalance and data scarcity in defect detection, enabling improved deep learning applications in 3D metrology.
Key Points
- Dataset includes 24,000 instances across 12 gear designs and 4 quality classes.
- Synthetic data generation helps mitigate class imbalance in defect detection.
- Each gear design has 500 instances, enhancing model training diversity.
- Data is generated using parametric CAD software with controlled perturbations.
- Open-source dataset supports deep learning-based 3D metrology research.
DeepSignal Analysis
What happened
The MFGNet-Gear dataset has been introduced as a synthetic 3D dataset aimed at enhancing quality inspection in manufacturing. It includes 24,000 paired polygon meshes and point clouds across 12 gear designs and 4 quality classes, addressing data scarcity and class imbalance in defect detection.
Key evidence
- MFGNet-Gear consists of 24,000 paired polygon meshes and point clouds, generated from 12 gear designs and 4 quality classes, with 500 instances for each design-quality combination.
- The dataset was created using parametric CAD software, with dimensional parameters adjusted by ±0.0254 mm and defect parameters sampled from specific distributions.
- MFGNet-Gear supports various applications such as part design classification, geometric defect detection, and representation learning, making it a valuable resource for deep learning in 3D metrology.
Why it matters
The introduction of MFGNet-Gear is significant as it addresses the challenges of obtaining labeled data for deep learning applications in manufacturing quality inspection. By providing a balanced and extensive dataset, it can potentially improve the performance of models used for defect detection, which is crucial in maintaining manufacturing quality.
Paper Resources
Source Excerpt
Quality control in smart manufacturing increasingly relies on data-driven methods, particularly deep learning, to automate the inspection of manufactured parts. Recent advances in three-dimensional (3D) metrology have enabled fine-scale assessment of dimensional accuracy, surface quality, and shape conformity. However, deep learning methods for point-cloud-based inspection require large volumes of labeled data covering part designs and defect types, which are costly and time-consuming to obtain.
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.