|
|
Foundations of Robotics
Seminar, October 25, 2006
Time
and Place | Seminar Abstract | Speaker
Biography | Presentation Slides | Speaker
Appointments
Fitting Joint Models to
Motion Capture Data
Lillian Chang
Smith Hall 100
Refreshments 4:15 pm
Talk 4:30 pm
Human motion analysis can be used to
synthesize captivating animation, evaluate biomechanical function, and inspire
humanoid robot design. A key component of movement analysis is the skeletal
model used to represent the body’s kinematics. Subject-specific models
are desirable for accurately describing anthropomorphic dimensions and
capturing individual style. In this talk, I will present techniques for fitting
joint models to motion capture data which have improved robustness over the
existing least squares methods. Two problems are considered: estimating the center of rotation of a ball joint and estimating the axis
of rotation of a hinge joint. For both joint models, careful selection of an
objective function with appropriate geometric properties is key to developing
intuitive methods which are robust to joint range of motion. Finally, I will
conclude with a discussion of the challenges associated with marker-based
approaches for human movement analysis.
|
Speaker Appointments |
For appointments, please contact Lillian
Chang (lillianc@andrew.cmu.edu)
The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.