In this project a new stereo-based 3D head tracking technique, based on scale-invariant feature transform (SIFT) features, that is robust to illumination changes is proposed. Also two major tracking techniques based on normal flow constraint (NFC) and 3D registration based method (ICP) is reviewed and compared against our own. A 3D head tracker is very important for many vision applications. The resulting tracker output parameters can be used to generate a stabilized view of the face that can be used as input to many existing 2D techniques such as facial expression analysis, lip reading, eye tracking and face recognition.
Our system can automatically initialize using a simple 2D face detector. The face detector is only sensitive to frontal heads therefore the initial pose of the head can be assumed to be aligned with the camera. SIFT extracts salient points from the intensity images and by matching them between frames. Together with the depth image and the matched features we obtain 3D correspondences. Using the unit quaternion method we recover the 3D motion parameters.