Head pose estimation using histogram of SIFT descriptors

Human head pose estimation is an important issue and a great challenge in many applications such as human-computer interaction, video conferencing and driver monitoring systems which has attracted many attentions in recent decades. In this paper we propose a novel method for human head pose estimation using Histogram of SIFT descriptors. Our method contains two phases: (1) preprocessing phase (2) obtaining Feature extraction set. Finally, for classification of our feature matrix using train and test samples, we take advantage of some well-known classifiers like: SVM, BayesNet and bagging via 10-fold cross validation technique to calculate the accuracy of our proposed algorithm. Results show that our proposed method outperforms previous methods in head pose estimation in terms of accuracy and efficiency.