Abstract
A novel feature extraction approach for fatigue expressions of vehicle drivers, which consists of Viola-Jones face detection algorithm and Gabor wavelet transform, was proposed. With features extracted from the fatigue expressions dataset created at Southeast University, holdout experiments on fatigue expressions classification are created using Multilayer Perceptron (MLP) classifier, compared with naive Bayes classifier, subspace classifier, and k-Nearest Neighbor (kNN) classifier. The results of holdout experiments show that MLP classifier offers the best classification performance than the other three classifiers. Among there predefined classes of driver's fatigue expressions, i.e., awake expressions, moderate fatigue expressions, and severe fatigue expressions, the class of severe fatigue expressions is the most difficult to classify. With MLP classifier, the classification accuracies of severe fatigue expressions are over 75% in holdout experiments, which shows the effectiveness of the proposed feature extraction method and the importance of MLP classifier in developing Driver Assistance Systems (DAS).
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call