English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Synthetic Data Generation Framework for Automated Quality Control in Gravure Printing

Forum topic · 小凯 · 2026-07-27

Summary

A paper on arXiv (2507.21739) by Korota Arsène Coulibaly, Mohamed Hamlich, and Khalid Hmli introduces a synthetic data generation framework for automating quality control in rotogravure printing. Because training robust deep learning defect detectors such as YOLO or Vision Transformers is hindered by the scarcity of real industrial defect images, the proposed pipeline automatically generates high-fidelity images of specific printing defects—creases, streaks, misregistration, and others—along with bounding boxes and annotations. To validate the approach, the authors generated a synthetic dataset of 7,533 images and trained the state-of-the-art object detection model RF-DETR on it. The trained model achieved 80.9% mean average precision (mAP) on real industrial test samples. The framework offers a zero-cost, rapid-deployment solution for defect detection on printing production lines without extensive manual data collection.

Overview

Field: Computer Vision Authors: Korota Arsène Coulibaly, Mohamed Hamlich, Khalid Hmali Published: 2025-07-27 arXiv: 2507.21739

Full Translation

Quality control in printing, particularly in rotogravure printing, still depends on slow, costly, and subjective manual inspection. Automated surface defect detection is critical for maintaining high-quality standards in rotogravure printing. Deep learning models offer prospects for automation. However, training robust deep learning models, such as YOLO or Vision Transformers, is heavily hindered by the extreme scarcity of real-world industrial defect images.

To overcome this limitation, this paper introduces a novel synthetic data generation framework tailored for rotogravure printing quality control. The proposed pipeline automatically generates high-fidelity images of specific printing defects (creases, streaks, misregistration, etc.) and outputs corresponding bounding boxes and annotations.

To validate the framework, the authors generated a synthetic dataset of 7,533 images and used it to train RF-DETR, a state-of-the-art object detection model. Experimental results show that the model trained on their synthetic data achieved 80.9% mean average precision (mAP) on real industrial test samples. The framework provides a zero-cost, rapid-deployment solution for automating defect detection on printing production lines, without extensive manual data collection.

Key Contributions

  • Synthetic data pipeline: automatic generation of high-fidelity defect images with bounding-box annotations for gravure printing
  • Defect coverage: handles creases, streaks, misregistration, and other common printing defects
  • Validation: RF-DETR trained on 7,533 synthetic images reaches 80.9% mAP on real industrial samples
  • Practical impact: eliminates the need for costly manual defect data collection on production lines

Tags

#computer-vision#synthetic-data#gravure-printing#quality-control#object-detection#rf-detr#defect-detection#deep-learning

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178503713