KaiChau Kung

Machine learning algorithm for classification of drowsy vs alert drivers

Started July 22, 2022

Abstract or project description

Every year, 100,000 police-reported car accidents are caused by drowsy driving in the US. 800 of those crashes results in death. Having an automated system for detecting drowsy driving would significantly reduce the number of accidents and save lives. In this project, my goal is to create a fast and reliable machine learning algorithm for the classification of drowsy vs alert drivers. First, we develop an image classification algorithm in python and train it on hundreds of images of individuals in all levels of alertness. Next, we test the algorithm and report the confusion matrix. This will give us the accuracy, the number of false positives and negatives, the sensitivity, and the specificity. Finally, based on these criteria, we will optimize the training process to increase accuracy and decrease the number of false positives and negatives during testing. The results of this project will contribute to safety on the road.