Recent Advances in Eye Tracking – Part 1

As part of trying to share our experiences and keep up-to-date with what the amazing eye tracking and research community is up to, in the past year we have participated in many interesting events. Some were quite technical and others delved into the methodological and theoretical aspects of eye tracking research.

In this series of articles I will go through some of the ideas and discussions presented during those events, starting with topics related to research methodology.

One question that many researchers asked was whether they should collect data from all available sensors and look for effects or systematic differences in the data later. During the talk of Prof. Demis Basso at the Eye Tracking and Biometric Systems conference, he suggested that the research question should be the drive for what data we collect and which sensors we choose to employ (Basso, 2017). The topic was brought up again during a panel discussion at a pre-conference workshop of the Design, Computation and Cognition conference in Lecco (2018). We argued that if the point of the study is to create a dataset for training machine learning algorithms then sure, dressing participants a suit of sensors is justified. However, if we were interested in studying gaze behavior in a visual search task, then eye tracking would suffice, or if we were interested in the temporal dynamics of brain activity during a decision task, an electroencephalograph would be the device of choice.

Another topic that was presented during the Eye Tracking and Biometric Systems meeting was the need for better communication between researchers and developers from different fields. Prof. Kristin Paetzold argued that psychologists should focus more on creating models that describe behavior, which in turn can be used by developers or engineers to design better interfaces or get a step further to humanoid robots that can perform more complicated tasks (Paetzold, 2017).

Other topics related to methodology were in the field of data visualization and visual analytics – the method of analyzing data and making inferences based on visualizations. During the Eye Tracking and Visualization conference in Warsaw (ETVIS, 2018) I got the pleasure to meet Dr. Michael Raschke, the co-founder of Blickshift, a great software that allows examining multiple recordings simultaneously by plotting different data streams on a shared timeline. It is a great way to detect certain patterns in the data. The software also offers sequence analysis for areas of interest and an interactive annotation tool to help you prepare your data for machine learning.

During ETVIS (2018) there were also some really inspiring presentations, including one on aggregated scan path visualization (Peysakhovich & Hurter, 2018ab). This innovative way of visualizing gaze data can help researchers get better insights just by looking at their data because it simplifies the typical scan path plot (or gaze plot) and it includes density and directionality of aggregated saccades in a single visualization. Their paper is available on this link to see some examples and how such graphs are calculated.

In the next part of this series I will get into a more technical aspect of eye tracking research, namely classifying different types of eye movements from raw eye tracking data.

Feel free to contact me if you have any questions or might need help dealing with your data for visualization or analysis.

Iyad Aldaqre

Data Scientist at SR Labs Srl

References

Basso, D. (2017) When are studies with biometric devices worth? The psychology standpoint. Presented at Eye tracking and biometric systems: breaking into industrial engineering, Bolzano-Bozen, December 7th 2017.

Paetzold, K. (2017) User experience and human factors as a fundamental driver for designers. Presented at Eye tracking and biometric systems: breaking into industrial engineering, Bolzano-Bozen, December 7th 2017.

Vsevolod Peysakhovich, V. & Hurter, C. (2018a) Intuitive Visualization Technique to Support Eye Tracking Data Analysis: A User-Study. In Chuang, L,. Burch, M. & Kurzhalz, K. (eds.), Proceedings of the 3rd Workshop on Eye Tracking and Visualization (ETVIS ’18), Warsaw, June 14 – 17 2018, ACM, New York, USA.

Peysakhovich, V. & Hurter, C. (2018b). Scan path visualization and comparison using visual aggregation techniques. Journal Of Eye Movement Research, 10(5). doi:http://dx.doi.org/10.16910/jemr.10.5.9