DRIVER DROWSINESS DETECTION USING CONVOLUTION NEURAL NETWORKS ALGORITHM

DRIVER DROWSINESS DETECTION USING CONVOLUTION NEURAL NETWORKS ALGORITHM

Lakshmi Siva Prasad¹, B. Venkata Lakshma Reddy², D. Ajay³, A. Jospin⁴

¹Student, blsprasad12@gmail.com

 ²Student, beeravallivenkatalakshmareddy@gmail.com

³Student, doppalapudiajay9848@gmail.com

⁴Student, amruthapudijospin@gmail.com

Dept. of E.C.E, R.V.R. and J.C College of Engineering

Abstract - Driving when fatigued is a severe problem that can result in tragic collisions. In this paper, we propose a convolutional neural network-based method for detecting driver drowsiness. Algorithm for neural networks (CNN). Real-time photos of the driver's face are captured by the proposed system using a camera, and these images are preprocessed before being fed into the CNN model. Based on the attributes gleaned from the face photos, the CNN model is trained to categories the driver's level of tiredness as either alert or drowsy. The results show how well the CNN algorithm performs in reliably identifying driver tiredness. A dataset of drivers in various levels of tiredness was used to evaluate the proposed system, and the results show the effectiveness of the CNN algorithm in identifying driver fatigue with accuracy. The suggested technology may be incorporated into already-installed car safety systems, adding a layer of protection to combat fatigued driving and enhance road safety.

Keywords- OpenCV, Keras, Convolution Neural Networks, Tensor flow, web camera and pygame.