Driver Identification System Using Finger Vein and YOLO Object Detection

Abstract

Biometrics used to identify people is one of the safest and most convenient identification methods. The hypoxic hemoglobin method can present a human finger vein image for driver identification through infrared ray irradiation. The finger vein image is captured system; with the finger vein image processed using contrast-limited adaptive histogram equalization (CLAHE) and the Gabor filter to obtain a clearer image. The YOLO object detection technology is used to drive identification. The proposed system can be divided into two parts. The first part is the training part, which processes images. A database generates the weight files for external testing. The second part is the testing system. When the driver's finger is placed in the designated photo area a photo will be taken. The photo will be processed and run on Raspberry Pi 4 together with the weight file to identify the driver. The experimental results show that when the database has a large amount of data, the yolov4-tiny-hy recognition rate is comparable to that of YOLOv4. The training time is greatly shortened.