[Dutch-chapter] September 9, Symposium on Observing and Generating Human Motion
Dennis Reidsma
d.reidsma at utwente.nl
Tue Aug 30 15:13:56 UTC 2011
Dear Reader,
it is our pleasure to invite you to the symposium "Observing and
Generating Human Motion", held on September 9 on the occasion of the PhD
defense of Herwin van Welbergen. Four internationally renown speakers
will present their work on cognition models and personality types for
virtual agents and monitoring and tracking of human motion.
Following the symposium, Herwin van Welbergen will give a short
introduction of the work that he did in the past few years, concerning
"Behavior Generation for Interpersonal Coordination with Virtual
Humans". Immediately thereafter, the official defense will take place.
You are kindly invited to attend the symposium as well as the
introduction and defense. Registration is not mandatory; however, if you
announce your intention to participate we can take it into account when
organizing coffee and additional copies of the thesis.
Date: September 9
Location: University of Twente
Program:
13:00 - 15:30 : Symposium, Room H327, Building Citadel
16:30 - 16:45 : Informal introduction by Herwin van Welbergen, Room WA4,
Building Waaier
16:45 : PhD Defense Herwin van Welbergen, Room WA4, Building Waaier
Speakers:
Michael Neff (US) -- Generation of personality types for virtual agents
Robby Tan (NL) -- Multiple people tracking from multiple cameras
Peter Veltink (NL) -- Monitoring daily-life physical interaction between
people and their environments
Stefan Kopp (DE) -- Models of human communication and cognition
Summaries of the talks:
===========
Michael Neff, University of California, US: generation of personality
types for virtual agents
A key goal in agent research is to be able to generate multi-modal
characters that can reflect a particular personality. Michael Neff will
talk about his recent work on non-verbal and verbal generation of
particular personality types. Based on psychological literature on the
relation between certain verbal and non-verbal behaviors, such as
self-adaptors and language variation, and personality models such as the
Big Five model, Michael Neff and his colleagues have experimented with
changing the perceived personality of a virtual agent.
===========
Robby Tan, Utrecht University: multiple people tracking from multiple
cameras
Multiple people tracking from multiple cameras benefits many
applications in computer vision. However, using the existing methods,
various problems such as inter-person occlusions still degrade the
position estimations. In our research, we attempt to solve the problems
by analyzing the view visibility and ranking the reliability of the cues
from 2D views. We combine the visibility with the smoothness constraints
into a probability framework, which offers a more flexible and robust
estimation. Aside from that, we also introduce the 2D reference lines to
estimate the 2D position of every person in the input images. These
lines are able to estimate more accurate and robust 2D positions. We
quantitatively evaluate our method by using both our own multiple-people
data set and a public data set. The evaluation and experimental results
on the standard data set show that our methods considerably improve the
accuracy.
===========
Peter Veltink, University of Twente: Monitoring daily-life physical
interaction between people and their environments
Continuous daily-life monitoring of the functional activities of stroke
survivors in their physical interaction with the environment is
essential for optimal guidance of rehabilitation therapy by medical
professionals and coaching of the patient. Such performance information
cannot be obtained with present monitoring systems. Human performance
and interaction with the environment is also important in other fields
like ergonomics and sports. It is the objective of our current research
to develop and validate sensing systems and algorithms for monitoring
interactions with the environment during daily-life. In the FreeMotion
project methods have been developed to monitor balance control using
shoes instrumented with inertial and force sensors. In our current STW
project PowerSensor, we develop methods to monitor power exchange
between the human body and the environment and dynamic characteristics
of the environment using a glove instrumented with inertial and force
sensing. In our new EU project INTERACTION, we develop and evaluate a
system to monitor daily physical interaction with the environment and
task performance after stroke. The system will be unobtrusively
integrated in clothing (e-textile), include fabric-based and distributed
inertial sensing, and provide telemonitoring and adaptive on-body
feedback capabilities. Telesupervision facilities will enable a clinical
expert at a distance to evaluate performance effectively, coach the
patient and influence training.
===========
Stefan Kopp, Bielefeld University, Germany: models of human
communication and cognition
The Sociable Agents Group at Bielefeld University develops technical
systems that can join humans naturally. Using 3D virtual humans or
humanoid robots, this research investigates how models of human
communication and cognition can be transferred to machines in order to
make them conversational, cooperative, adaptive and companionable. I
will give an overview of ongoing work on (1) modeling speech and gesture
and studying their effects, (2) modeling feedback understanding to
unveil the interlocutor's degree of understanding and responding by
adaptation of own linguistic behavior, (3) modeling online learning and
recognition of nonverbal behavior.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.digra.org/pipermail/dutch-chapter/attachments/20110830/49e7de88/attachment.html
More information about the Dutch-chapter
mailing list