Purchase this article with an account.
Natalia Bogdan, Robert Allison, Rajaraman Suryakumar; Infrared Tracking of the Near Triad. Journal of Vision 2010;10(7):507. doi: https://doi.org/10.1167/10.7.507.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The oculomotor response when viewing a near target is characterized by ‘the near triad’: pupil miosis (constriction), binocular convergence and increased accommodation. Most existing eye-tracking systems lack the ability to measure all three of these parameters and are usually specialized to handle only one. Systems that can measure the complete near triad suffer from slow measurement rates, off-line analysis or are cumbersome and inconvenient to use. Singular specialized systems are usually combined ad-hoc but such systems are often complex in architecture and suffer severe limitations in runtime. We describe a video-based eye tracking system based on eccentric photorefraction that allows for remote, high-speed measurement of all three components of the near triad. This provides for precise, simultaneous measurement of oculomotor dynamics as well as having the benefit of being safe and non-intrusive. An extended infrared source illuminated the subject's eye. The corneal reflex and ‘bright pupil’ reflections of this source were imaged by an infrared sensitive camera and used to track gaze direction and pupil diameter. Such eccentric illumination combined with a knife-edge camera aperture allowed the accommodative state of the eye to be estimated from measurements of the gradient of image intensity across the pupil. Real-time measurements are facilitated by detection of Purkinje images to define areas of interest for each pupil followed by pupil edge detection and fitting to an ellipse model. Once the pupils are located, data about the brightness profile, diameter, corneal reflex and pupil center are extracted and processed to calculate the near triad. The system will be used in ongoing experiments assessing the role of oculomotor cues in perception of motion in depth.
This PDF is available to Subscribers Only