In this paper, we present a real time algorithm for mobile robots to track human faces and estimate face poses accurately, even when humans move freely and far away from the camera or go through different illumination conditions in uncontrolled environments. We combine the algorithm of an adaptive correlation filter with a Viola-Jones object detection to track the face as well as the facial features including the two external eye corners and the nose. These facial features provide geometric cues to estimate the face pose robustly. In our method, the depth information from a Microsoft Kinect camera is used to estimate the face size and improve the performance of tracking facial features. Our method is shown to be robust and fast in uncontrolled environments.
[pdf]
@inproceedings{voduc2013,
author = {My Vo Duc and Andreas Zell},
title = {Real Time Face Tracking and Pose Estimation Using an Adaptive Correlation
Filter for Human-Robot Interaction},
booktitle = {European Conference on Mobile Robots (ECMR 2013) (Oral)},
year = {2013},
address = {Barcelona, Catalonia, Spain},
month = september,
abstract = {In this paper, we present a real time algorithm for mobile robots
to track human faces and estimate face poses accurately, even when
humans move freely and far away from the camera or go through different
illumination conditions in uncontrolled environments. We combine
the algorithm of an adaptive correlation filter with a Viola-Jones
object detection to track the face as well as the facial features
including the two external eye corners and the nose. These facial
features provide geometric cues to estimate the face pose robustly.
In our method, the depth information from a Microsoft Kinect camera
is used to estimate the face size and improve the performance of
tracking facial features. Our method is shown to be robust and fast
in uncontrolled environments.},
url = {http://www.cogsys.cs.uni-tuebingen.de/publikationen/2013/voduc2013.pdf}
}