Seminar Details
2016-02-09 (14:00) : Riemannian geometry applied to EEG-based cerebral interfaces: Challenges and opportunities
At Euler building (A002)
Organized by Mathematical Engineering
Speaker :
Sylvain Chevallier (Laboratoire d'Ingénierie des Systèmes de Versailles
)
Abstract :
Brain computer interfaces (BCI) allow to interact with a computer without relying on the user's motor capabilities. In rehabilitation and assistive technologies, BCI offer promising solutions to compensate for physical disabilities. A common choice for BCI is to rely on electroencephalogram (EEG) recording for detecting specific brain waves. These brain signals could be recorded with lightweight equipment and have a good temporal resolution but are plagued by several sources of noise (electronic, electrical and biological). Most of the state-of-the-art approaches deal with covariance matrices to reduce the effect of these noise. Indeed, Riemannian geometry has provided a substantial framework for developing new algorithms. Most notably, a straightforward algorithm such as Minimum Distance to Mean yields competitive results when applied with a Riemannian distance. Combined with other approaches, algorithms based on Riemannian geometry have won several machine learning
challenges, outperforming Euclidean-based algorithms. After introducing the basics of electroencephalography and brain computer interface, this talk will explore the benefits brought by the Riemannian distances and their properties.