An Automated Detection Model of Vehicle Identification Number Using YOLOv5

Abstract

The identification plate of vehicles is not an effective way of identifying vehicles because it can be stolen, removed, or altered. Therefore, the present research suggests utilizing the advantages of using the vehicle identification number (VIN) instead of vehicle plates in determining vehicle identities.
This paper proposes an automated detection model of vehicle identification number using YOLOv5.
The detection model was tested using Roboflow dataset, which consists of 2797 images of quality 256. The experimental results reach 80.3% in mAP 0.5:0.95 and 99.4% in mAP 0.5.