INTEGRATING COMPUTER VISION INTO EQUIPMENT MAINTENANCE AT CHU-UFPA
This study applies computer vision techniques to detect defects in printed circuit boards of multiparameter monitors and pulmonary ventilators, in order to reduce corrective maintenance. Using transfer learning and data augmentation, the model achieved 77.77% accuracy, despite limited real data.
Institutions
- Universidade Federal do Pará (BR)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2025-10-16
- DOI
- https://doi.org/10.5281/zenodo.17363837
- Primary Topic
- Industrial Vision Systems and Defect Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00