A group of international scientists including an Indian-origin graduate student is developing which could let the users control their smartphone through eye movements to open apps, play games and do other smarphone related stuff.
The group consists scientists from USA’s University of Georgia, Massachusetts Institute of Technology (MIT) and Germany’s Max Planck Institute for Informatics. And according to the MIT Technology Review report, till now the scientists have been able to train the software to identify in which way a person is looking, with an accuracy of about a centimetre on a mobile phone and 1.7 centimetres on a tablet.
To create the software, the scientists developed an app named as GazeCapture which accumulated data about how people look at their handsets in different environments outside the confines of a lab. And as per the study co-author Aditya Khosla of MIT, the software’s accuracy will improve gradually, with more sample size & data.
With the help of the handset’s front camera, GazeCapture recorded the user’s gaze, as they were shown pulsating dots on the phone’s screen. Then to ensure that the users are giving attention, they were shown a dot with an “R” or “L” inside it and asked to tap the right or left side of the screen in response.
Now the information captured by GazeCapture used to train another software, dubbed as iTracker, made by Apple and can even run on your iPhone. As the smartphone’s front camera captures your face, iTracker considers the other factors such as direction and position of your eyes and head to figure out whether the user’s gaze is focused on the phone’s screen or not.
The scientists have considered a sample size of nearly 1500 people so far, and they are trying to get data from 10,000 people who have used the GazeCapture app. Khosla said, if they manage to get the sample of that size, it will help them to reduce iTracker’s error rate to half a centimetre, a mammoth improvement for a range of eye-tracking applications.
The research team also presented their study results at the IEEE Conference on Computer Vision and Pattern Recognition in Seattle, Washington, recently. Khosla also added, apart from day to day usage, the eye-tracking software could be used for medical diagnoses purpose too, especially to diagnose those patients who are suffering from schizophrenia and concussions.
