NIR Camera for Gaze Tracking

Recommended reading in Optical Engineering.

01 October 2016
Eddie Jacobs

Are you looking at me?

The question is not just an iconic line from the 1970s era movie “Taxi Driver.” It’s a legitimate research question and the subject of a recent Optical Engineering paper, “Discriminating between intentional and unintentional gaze fixation using multimodal-based fuzzy logic algorithm for gaze tracking system with NIR camera sensor.”

The open-access article is authored by Rizwan Ali Naqvi and Kang Ryoung Park from Dongguk University (Republic of Korea) and appears in the June 2016 issue.

Experimental setup of a gaze-tracking system. An NIR illuminator is positioned below the camera to best detect changes in the size of the pupil.

Detecting gaze fixation allows for a variety of hands-free human-computer interfacing options. These can be used to assist the handicapped in communication and navigation.

The authors review various methods for discriminating whether gaze fixation is intentional or unintentional. They then introduce a method based on the detection of pupil size changes, which indicate a cognitive task is being performed.

They couple this feature with the dwell time of the gaze. An algorithm combining standard machine-learning techniques with fuzzy logic is then used to discriminate intentional and unintentional gaze fixations.

The algorithm seems to perform well as demonstrated in the accompanying experimental verification. This is all done using off-the-shelf imaging and computational components.

The system described in this paper could become part of the next generation of human-computer interfaces.

With a gaze and a thought, we might place a call, order from a menu, or even hail a taxi.


SPIE Fellow Eddie L. Jacobs of University of Memphis is a member of the Optical Engineering editorial board
SPIE Fellow Eddie L. Jacobs of University of Memphis is a member of the Optical Engineering editorial board.

Recent News
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research