Eye movement as a biometric identifier progresses with new research; market growing
New research has come out looking at how eye movements can be used for biometric identification.
A team of researchers at the Silesian University of Technology in Poland found eye behaviors unique to 24 study participants. The behaviors were gleaned from how volunteers performed in experiments in which they had to follow a point racing around 29 locations on a screen.
The researchers evaluated four classification methods in tests with training datasets generated in different ways, and found that the random forest method offered the best overall performance. Next experimental steps are set out in the paper’s conclusion.
This is an arcane field of research that stretches back at least to 2005, with significant new work appearing in 2012, with increasingly great details being observed in decreasing fractions of a second.
Among the aspects of eye movement studied in the new report were eye velocity and acceleration as subjects chased after the points.
Scientists also looked at saccades, rapid, sudden movements during which no information is collected by the brain, and fixations, when eyes focus on an object. There is data to be gathered then, too, because eyes are not motionless even when staring at something. Those small movements, described as “gentle,” can be biometric identifiers.
The eye-tracking economic segment, which encompasses biometric security as well as gaming, automaking and health care roles, reportedly is growing. An eye-tracking virtual-reality headset that tracked eye movement, or signals, was introduced by Japanese startup Fove in 2016.
A study last October by Grand View Research projected a compound annual growth rate from 2018 to 2025 of 26.3 percent. At the end of that period, according to the firm, global eye tracking will generate $1.75 billion.