Drowsiness Detection Using Convolutional Neural Networks

• Published in: 2023 International Conference on Research Methodologies in Knowledge Management, Artificial Intelligence and Telecommunication Engineering (RMKMATE) pp. 1 - 6
• Main Authors: Samuel, Prithi, Pascal, Dennis, M, Kiruthika, Manoj, Abhijjith
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2023
• Subjects: Biomedical monitoring; CNN; drowsiness detection; eye aspect ratio; Learning systems; machine learning; Magnetic heads; Roads; Streaming media; Visualization; Wheels
• DOI: 10.1109/RMKMATE59243.2023.10369658

Growing number of jobs in today's economy demand sustained concentration. Drivers need to keep a close eye on the road so they can react swiftly to unforeseen circumstances. Driver weariness frequently contributes immediately to a number of traffic accidents. In this way, it is envisioned that frameworks would be developed to identify and alert drivers to their severe psychophysical conditions, perhaps reducing the incidence of accidents involving exhaustion. However, there are a number of challenges in developing these devices that are related to accurate and timely recognition of a rider's signs of exhaustion. Utilising the vision-based approach to detect driver fatigue and slow down or stop the vehicle when it is detected is one of the specialised outcomes that could be achieved. Here, we assess the visual framework to detect driver drowsiness. Various approaches have been proposed for drowsiness detection, including physiological measures, behavioural measures, and machine learning-based methods. Physiological measures involve the measurement of body signals, such as heart rate variability, eye movement, and brain activity. Behavioural measures include changes in driving performance, such as lane deviation and steering wheel movement. Machine learning-based methods involve the use of artificial intelligence techniques, such as deep learning and neural networks, to detect drowsiness