An eigenface approach for estimating driver pose
Paul B. Watta, Niketa Gandhi, Sridhar Lakshmanan · 2002
In this paper, we present a non-parametric method which can be used to analyze facial video data of an automobile driver as he or she drives the vehicle. Each frame in the video sequence is classified using an eigenface representation. A database of face pose images is constructed, and experimental results are given which measure the performance of the method on a large test set. Variations in the performance as the number of faces used to train the classifier, as well as the number of eigen coefficients in the representation are varied, are also reported.