README.md

Assessing Visual Attention Using Eye Tracking (Independent Research Project)

Author

Ioana Munteanu

Mentor

Stephen J. Guy

Abstract

The scope of this research is to determine if there is any correlation between the level of experience of surgeons and their visual attention while performing surgeries. Data was gathered by Dr. Sofia Lyford-Pike and her team. 18 surgeons at the University of Minnesota performed surgeries while using Pupil Lab’s mobile eye-tracking glasses. By analyzing the position, velocity, and acceleration of the pupils, a percentage of focus was computed. It was determined with a confidence of 95% that a surgeon is focused if their pupils' velocity is under 3 pixel/s and acceleration under 500 pixel/s². Using different statistical models, it was determined that senior surgeons are more visually focused than junior surgeons.

Poster

To get the PDF version, click here.

Programming languages, tools, and frameworks

• Python
• Matplotlib
• NumPy
• R
• Excel

Files

• Assessing Visual Attention Using Eye Tracking.pdf: poster
• percentage_focus.py: algorithm sed to determine the percentage of focus, as described in "Velocity as a measure of focus"
• plot.py: helper functions used to graph the data
• results_table.csv: table containing data about residency year, average velocity during the scrub-in and during the actual surgery, and the difference between the two velocities