MPI-INF Logo
Campus Event Calendar

Event Entry

What and Who

Smooth pursuit eye movement classification for clinical diagnosis and everyday applications

Michael Dorr
TU München
Talk

Michael Dorr studied Computer Science at the University of Luebeck, Germany. After receiving a doctorate in Engineering in 2010, he moved to Boston to become a postdoctoral fellow in the Department of Ophthalmology, Harvard Medical School. In 2014, Michael joined Technical University Munich as a TUM Junior Fellow to head the research group "Visual Efficient Sensing for the Perception-Action Loop", funded by the Elite Network Bavaria. He is also co-founder of Adaptive Sensory Technology, Inc., a medical device spin-off company.
AG 1, AG 2, AG 3, AG 4, AG 5, RG1, SWS, MMCI  
Public Audience
English

Date, Time and Location

Monday, 1 August 2016
15:00
60 Minutes
E1 4
024
Saarbrücken

Abstract

In order to reduce the computational burden of visual information processing, humans and many other animals have evolved a variable-resolution retina that is directed to informational regions of the visual input by means of several eye movements per second. As a consequence, we can make inferences about an observer by analysing their gaze behaviour. Smooth pursuit eye movements, which can be regarded as fixations on moving targets, constitute an important aspect of this behaviour, but to date have received relatively little attention in eye movement research because of the predominant use of static stimuli where they cannot occur.
Here, we will present recent results obtained with a novel approach to reliably detect smooth pursuit episodes in noisy gaze traces from many observers. We will show that smooth pursuit eye movements are a frequent occurrence when observers watch dynamic natural scenes, and how this behaviour may be impaired in patients with mental disorders such as schizophrenia. Furthermore, we will describe a new human-computer interaction method that uses smooth pursuit eye movements as an input modality and that shows a greatly reduced error rate compared
to dwell time, the current gold standard in gaze-based interfaces.

Contact

Connie Balzert
--email hidden
passcode not visible
logged in users only

Connie Balzert, 06/20/2016 12:53 -- Created document.