Automated defect recognition of castings defects using neural networks Articles uri icon

publication date

  • March 2022

start page

  • 11-1

end page

  • 11-15


  • 1


  • 41

International Standard Serial Number (ISSN)

  • 0195-9298

Electronic International Standard Serial Number (EISSN)

  • 1573-4862


  • Industrial X-ray analysis is common in aerospace, automotive or nuclear industries where structural integrity of some parts needs to be guaranteed. However, the interpretation of radiographic images is sometimes difficult and may lead to two experts disagree on defect classification. The automated defect recognition (ADR) system presented herein will reduce the analysis time and will also help reducing the subjective interpretation of the defects while increasing the reliability of the human inspector. Our convolutional neural network (CNN) model achieves 0.942 mAP@IoU = 0.50, which is considered as similar to expected human performance, when applied to an automotive aluminium castings dataset (GDXray). On an industrial environment, its inference time is less than 400 ms per 16 GB DICOM image (16 bits), so it can be installed on production facilities with no impact on delivery time. In addition, an ablation study of the main hyper-parameters to optimise model accuracy from the initial baseline result of 0.75 mAP up to 0.942 mAP, was also conducted.


  • Computer Science
  • Mechanical Engineering
  • Robotics and Industrial Informatics
  • Telecommunications


  • x-rays; castings defects; automated inspection; adr; convolutional neural network; retinanet