Hybrid Approach for Driver Drowsiness Detection

In today’s world Driver drowsiness is a major cause of road accidents, necessitating the development of robust and accurate drowsiness detection systems. Traditional single-feature approaches, such as blink rate and face detection often lack reliability due to environmental variations and individual differences. This paper proposed a hybrid approach with multiple features drowsiness detection. The paper explored Media Pipe and Dlib's 68-face landmark detectors and verify the accuracy. Benchmark datasets such as UTA- RLDD, YawDD and DROZY are used to analyse the performance of the proposed technique using performance metrics like accuracy, precision, recall and F1 score. This paper present comparative analysis of proposed hybrid technique with existing approach. The integration of the proposed technique with advanced driver assistance system (ADAS) and autonomous vehicles will ensure enhancement of the road safety.