Visual Feature Based Data Reduction

PI: Ingrid Hotz (LiU), co-PI: Bo Bernhardsson (LU); with Anders Ynnerman (LiU), Michael Dogget (LU), Jonas Unger (LiU), Bo Bernhardsson (LU)

Developing Brain-Computer interfaces (BCIs) is a challenging goal that, however, is becoming more and more realistic. Brain imaging and measurement methods provide valuable data to gain the necessary knowledge about the complex functionality of the brain.  The goal of the project is to make such brain data accessible for effective exploitation in medical and technological applications. The underlying concept is to develop sparse representations that can be used to guide a visual exploration process. We plan to approach this challenge by combining geometric and topological methods for dimension reduction, with learning methods to classify the obtained features, and modern interaction and rendering facilities to communicate the results.

Project number: B12