\section{Introduction}
\label{sec:intro}

Person identification has become a valuable asset, whether for means of
authentication, personalization, or other applications.  Previous work revolves
around either physiological biometrics, such as face recognition, or behavioral
biometrics such as gait recognition.  In this paper, we propose using
skeletal measurements as a new physiological biometric for recognition.

In recent years, advances in range cameras have given us access to increasingly
accurate real-time depth imaging.  Furthermore, the low-cost and widely
available Kinect~\cite{kinect} has brought range imaging to the masses.  In
parallel, the automatic detection of body parts from depth images has led to
real-time skeleton mapping.  As the resolution and accuracy of range cameras
improve, so will the accuracy and precision of skeleton mapping algorithms.  

In this paper we show that skeleton mapping is accurate and unique enough in
individuals to be used for person recognition.  We make the following two
contributions.  First, we show that ground truth skeleton measurements can
uniquely identify a person.  We model how the accuracy of skeleton recognition
decreases as simulated error increases, and find it is still possible to use
for recognition.  Second, we evaluate our hypothesis using real-world data
collected with the Kinect.  Our results show that skeleton recognition performs
well, particularly in situations where face recognition cannot be performed.

Much of the prior work in person recognition focuses on data gathered from
other sensors, such as face recognition with color cameras and voice
recognition with microphones.  In the realm of depth imaging, most of the work
surrounds behavioral recognition, continuing work in gait recognition.  The
Xbox 360~\cite{} does use the height inferred from the Kinect as part of its
user identification algorithm, albeit in addition to other attributes including
face recognition.

The paper is organized as follows.  First we discuss prior methods of person
identification, in addition to the advances in the technologies pertaining to
skeleton mapping (Section~\ref{sec:related}).  Next we use a dataset of actual
skeletal measurements to show that identification by skeleton is feasible, even
when we simulate the error of measuring skeletons with a Kinect
(Section~\ref{sec:}).  Finally, we Lastly, we collect skeleton data with a
Kinect in an uncontrolled setting and we apply preprocessing and classification
algorithms to this dataset (Section~\ref{sec:}).  We evaluate the performance
of skeleton recognition with varying group size and compare it to face
recognition. We conclude by discussing challenges working with the Kinect and
future work (Section~\ref{sec:conclusion}).