The PUPILLO tracker: next generation eye tracking

Introduction

Eye-tracking is one of the most valuable tools for measuring human behaviour, applicable to multiple populations and stages of development, and their use has enjoyed an exponential increase both within and outside of research. Pupil size, measurable with eye-trackers, provides a valuable index of a wide range of cognitive functions.

Problem Statement

However, pupillometry has significant untapped potential because current commercial eye-trackers are unable to provide reliable indications of cognitively driven pupil changes. This results in:

• A restricted design space for research
• Large loss of data
• Often unreliable measurement

Proposed Solution

We propose a novel architecture of remote eye-tracker for behavioural research, conceived to address both the problem of large data loss and the problem of reliability of pupil size changes. Our solution addresses these pains at both the hardware and software levels, with a series of innovations including:

1. An identical offline mode to retrieve otherwise lost data
2. Autofocus and robotic tracking to accommodate much greater participant movement

Results

In pilot experiments with infants, we have increased the proportion of usable data from about 50% with the current state of the art to close to 100%.

Future Development

We are currently developing procedures and algorithms to compensate for the light reflex so that pupil dynamics data is deconfounded, and acquisition is possible over a much larger span of stimuli and situations.

Conclusion

PUPILLO aims to become the next generation eye-tracker, providing the research and medical communities and beyond with a cost-effective, easy-to-use, and customizable solution to fully exploit the potential of gaze tracking and pupillometry.