Real time 3D head pose estimation: Recent achievements and future challenges

Abstract

Most automatic face recognition algorithms try to normalize facial images in order to remove variations caused by anything but the identity of the person. Lighting conditions being less problematic since the introduction of reliable and affordable depth sensors, head pose is the other great source of un-desired variations in facial images. In this paper, we describe recent state-of-the-art methods for real time head pose estimation from depth data, present available databases, and discuss open problems to be addressed by future research.