teaching

:: master theses (6 months fulltime)

• Biometric interface for authentication using eye movements from EOG
  Marcel Breu, 2009 [Description]
  
  
  

  Goal
  Development and evaluation of a biometric authentication system based on eye movement analysis. Eye movements were recorded using a wearable electrooculographic (EOG) system.
  
  Method
  Evaluation of different visual stimuli, eye movement features, as well as feature selection and classification algorithms. Development of a prototype authentication system used in a user-study with 10 participants.
• Adaptive EOG signal processing on DSP
  Roman Suter, 2007 [Description]
  
  
  

  Goal
  Development of algorithms for embedded real-time processing of EOG signals on a wearable sensor. Particular emphasis was on methods for adaptive removal of EOG signal artefacts caused by physical activity.
  
  Method
  Implementation of algorithms for saccade detection, blink detection/removal, and baseline drift reduction in embedded C. Evaluation of a method for walking detection and adaptive filtering of EOG signals in a user study.
• Development of an integrated wearable EOG-based eye-tracker
  Roberto Celli, Christian Meyer, 2007 [Description]
  
  
  

  Goal
  Design and implementation of an integrated wearable eye-tracker based on Electrooculography (EOG) from scratch. Desired features include unobtrusive data collection, real-time signal processing, onboard storage and wireless data transmission.
  
  Method
  Evaluation of several possible designs, hardware development, manufacturing, and firmware/GUI programming. The prototype implemented as goggles was evaluated in an HCI user study with 11 participants.

:: semester theses (3 months fulltime)

• Long-Term Eye Movement Analysis Using Wearable Electrooculography
  Michael Hardegger, 2009 [Description]
  
  
  

  Goal
  Development and evaluation of algorithms for automatic segmentation of long-term eye movement data recorded using a wearable electrooculographic (EOG) system.
  
  Method
  Recording of long-term EOG data from eight participants in a naturalistic office scenario. Development of eye movement features and evaluation of algorithms using these features for segmentation.
• Online configuration and calibration for wearable EOG goggles
  Dominik Kamm, 2009 [Description]
  
  
  

  Goal
  Development and implementation of an automatic online configuration procedure for the wearable EOG-based eye tracker. The procedure needed to cover both the saccade as well as the blink detection algorithms.
  
  Method
  Development of an appropriate configuration sequence. Implementation and evaluation of different algorithms for automatic thresholding and configuration in MATLAB and C++.
• Text-entry interface using continuous eye gestures from EOG
  Stefan Dahinden, Stephan Lütolf, 2008 [Description]
  
  
  

  Goal
  Design and implementation of a continuous text-entry interface using discrete eye gestures recognised from electrooculographic (EOG) signals similar to Dasher for gaze-based input.
  
  Method
  Analysis of Dasher's input characteristics, development of a prototype interface using eye gestures for text-entry, and evaluation of the interface in a user study with eight participants.
• Automatic artefact compensation in EOG signals
  Philip Herter, Martin Wirz, 2007 [Description]
  
  
  

  Goal
  Implementation and evaluation of algorithms for EOG signal processing: saccade and blink detection and noise, baseline drift and eye blink artefact removal.
  
  Method
  Performance evaluation of a variety of algorithms on artificial and real-world EOG data. Development of new algorithms for eye tracker calibration and integration into an artefact compensation framework (ACF).

www.andreas-bulling.de | © 2006-2010 Andreas Bulling
last update: 02 March 2010