Abstract:
The dangers of distracted driving to the safety of the road are grave. In this study, a method for
real-time face and eye identification of drivers is proposed. It combines machine learning with a
convolutional neural network (CNN) approach. The recommended methodology includes data
collection and preprocessing, CNN model training, and the use of machine learning methods. To
ensure high-quality data, the approach begins by gathering a diverse collection of face and ocular
pictures, which are then preprocessed and annotated. The dataset is then evaluated using EDA
techniques to identify potential biases, highlight significant facial characteristics, and correct data
imbalances. The basis of this suggested system, which makes use of the OpenCV library, is the
analysis of face photographs, which allows us to predict the dynamics of tiredness or dizziness or
to limit the number of traffic accidents. This proposed system can also be used to evaluate the
effects of drowsiness or drowsy warnings under different driver conditions. We will accumulate
the results, and then we can use them in our system to reduce the number of cases, track the cases,
and further improve the system. This will play a very important role in saving lives and reducing
deaths and accidents worldwide.
