Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Human Pose Reconstruction from IMUs
We offer several projects to investigate inertial-based motion capture system for full-body pose reconstructions.
Keywords: IMU, Motion Capture, Pose Estimation
An inertial measurement unit (IMU) is a sensor that measures its own orientation and acceleration. Modern IMUs are microelectromechanical systems (MEMs) equipped with at least a gyroscope and an accelerometer. Typically, they also contain a magnetometer to sense the earth-magnetic field, which in turn stabilizes orientation measurements.
With respect to motion capturing, IMUs overcome one of the biggest limitations of optical tracking systems, which is the line-of-sight requirement. Additionally, they also allow for much more mobile capture setups. Hence, inertial-based motion capture systems have become strong competitors for optical tracking methods.
Nevertheless, using IMUs for motion capturing comes with some problems that open up a rich space for research projects. In our lab we are interested in investigating:
• Using a low sensor count for pose reconstruction, such as done here https://ait.ethz.ch/projects/2018/dip/
• Modelling of IMU characteristics to enable more efficient learning from its measurements.
• Improving 3D position estimation from IMUs alone.
If you are interested in working with IMUs in the context of body pose estimation, please reach out to discuss possible projects in more detail. We are also open to discuss and support your own ideas.
An inertial measurement unit (IMU) is a sensor that measures its own orientation and acceleration. Modern IMUs are microelectromechanical systems (MEMs) equipped with at least a gyroscope and an accelerometer. Typically, they also contain a magnetometer to sense the earth-magnetic field, which in turn stabilizes orientation measurements.
With respect to motion capturing, IMUs overcome one of the biggest limitations of optical tracking systems, which is the line-of-sight requirement. Additionally, they also allow for much more mobile capture setups. Hence, inertial-based motion capture systems have become strong competitors for optical tracking methods.
Nevertheless, using IMUs for motion capturing comes with some problems that open up a rich space for research projects. In our lab we are interested in investigating: • Using a low sensor count for pose reconstruction, such as done here https://ait.ethz.ch/projects/2018/dip/ • Modelling of IMU characteristics to enable more efficient learning from its measurements. • Improving 3D position estimation from IMUs alone.
If you are interested in working with IMUs in the context of body pose estimation, please reach out to discuss possible projects in more detail. We are also open to discuss and support your own ideas.