Estimation of Pose Parameters of Vehicle Drivers Based on Facial Feature Triangle
Zhang Wan-zh · Journal of the China Railway Society · 2013
A method of pose parameters estimation for vehicle drivers based on the facial feature triangle was proposed.Firstly,we used the Adaboost algorithm and expanded Haar features to train the classifier and locate eyes.Secondly,we extracted the Hcomponent of the mouth area to locate the mouth accurately on the basis of eye location.Finally,we established the facial feature triangle and put forward three formulas of head postures and estimated pose parameters on the basis of position changes of the facial feature triangle.Experiments show that three postures can describe the head accurately and the method only needs the coordinates of the mouth and eyes in the current image.The method is simple and fast.