Currently, Khosla says, their training set includes examples of gaze patterns from 1,500 mobile-device users.
Previously, the largest data sets used to train experimental eye-tracking systems had topped out at about 50 users.
To assemble data sets, "most other groups tend to call people into the lab," Khosla says. "It's really hard to scale that up. Calling 50 people in itself is already a fairly tedious process. But we realised we could do this through crowdsourcing,” he added.
In the paper, the researchers report an initial round of experiments, using training data drawn from 800 mobile-device users.
On that basis, they were able to get the system's margin of error down to 1.5 centimetres, a twofold improvement over previous experimental systems. The researchers recruited application users through Amazon's Mechanical Turk crowdsourcing site and paid them a small fee for each successfully executed tap. The data set contains, on average, 1,600 images for each user.
The team from MIT's Computer Science and Artificial Intelligence Laboratory and the University of Georgia described their new system in a paper set to presented at the "Computer Vision and Pattern Recognition" conference in Las Vegas on June 28.