A comparison of different calibration strategies for webcam-based eye-tracking.

What is it about?

We design and compare different calibration strategies and apply them for deep-learning based webcam eye-tracking in an online experiment. The results help understand the effect of calibration data size, frequency of calibration and effect of calibration task used on the final gaze prediction accuracy. We use the standard fixation accuracy measure to evaluate gaze performance.

Featured Image

Photo by Ion Fet on Unsplash

Photo by Ion Fet on Unsplash

Why is it important?

Webcam based eye-tracking allows researchers to sample global populations and replicate results cross-culturally, in an economical and time-efficient manner. These methods have to deal with the challenges of hardware limitations on temporal and spatial resolution, added noise in collected data due to uncontrolled environments, accurate estimation of physical parameters for remote calibration procedures etc. Our study aims to address these challenges and suggest suitable measures to collect accurate and reliable eye-tracking data during online studies.