Manufacturing Defect Detection Using Convolutional Neural Networks on Industrial Images

Authors

  • Ahmad Fauzan Audhighasu University Author
  • Putri Rahmadani Sepuluh Nopember Institute of Technology image/svg+xml Author

DOI:

https://doi.org/10.5555/audhighasu.v1i2.7

Keywords:

computer vision, CNN, defect detection, EfficientNet, industry 4.0

Abstract

Manual visual inspection is prone to human error. This study applies EfficientNet-B3 to detect surface defects on electronic products. A private dataset of 8,200 images covers six defect classes. The model achieves 0.91 mAP and runs on a Jetson Nano edge device at 22 FPS.

Readership pulse

How this article is used

Live abstract views and file downloads for this article.

Views 310 Abstract views
Downloads 219 File downloads
Total 529 All interactions

6-month trend

Download mix

Author Biographies

  • Ahmad Fauzan, Audhighasu University
    Peneliti machine learning dan visi komputer.
  • Putri Rahmadani, Sepuluh Nopember Institute of Technology
    Dosen teknik informatika.
Cover of Audhighasu Journal of Informatics Vol. 1 No. 2 (2024) — IT Innovation

Downloads

Published

2024-12-20

Issue

Vol. 1 No. 2 (2024): Vol. 1 No. 2 (2024): Information Technology Innovation

Section

Artikel

License

Copyright (c) 2024 Audhighasu Journal of Informatics

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.