Inexpensive optical device enables determination of eyeglass prescriptions

Cambridge, MA-- A team at MIT’s Media Lab has come up with a quick, easy and inexpensive method for determining eyeglass prescriptions. Based on mobile phones fitted with lenses, the approach is especially suitable for remote, developing-world locations.

The team is preparing to conduct clinical trials, but preliminary testing with about 20 people, and objective tests using camera lenses, have shown that it can achieve results comparable to the standard aberrometer test.

In its simplest form, the test can be carried out using a small, plastic device clipped onto the front of a cell phone’s screen. The patient looks into a small lens, and presses the phone’s arrow keys until sets of parallel green and red lines just overlap. This is repeated eight times, with the lines at different angles, for each eye. The whole process takes less than two minutes, at which point software loaded onto the phone provides the prescription data. The device is described in a paper by MIT Media Lab Associate Professor Ramesh Raskar, Visiting Professor Manuel Oliveira, and Media Lab student Vitor Pamplona (lead author of the paper) and post-doctoral research associate Ankit Mohan, that will be presented in late July at the annual computer-graphics conference SIGGRAPH.

The device uses an optical system derived from one some team members developed last year as a way of producing tiny barcodes (called Bokode) that could provide a large amount of information. Raskar explains that he had demonstrated that barcode device to many people, but when he showed it to his wife, she had trouble seeing its patterns. He quickly realized that others he had shown it to had been wearing their glasses or contact lenses, but his wife had been looking into it directly and it had revealed the imperfections in her vision.

The prototype system Raskar and his students developed as a result of that insight has an array of tiny lenses and a grid of pinholes that, combined with the software on the phone, “forces the user to focus at different depths” so the eye’s focusing ability can be measured. Essentially, Raskar explains, the test works by transforming any blurriness produced by aberrations in the eye into an array of separate lines or dots instead of a fuzzy blob, which makes it easier for the user to identify the discrepancy clearly. Rather than estimating which of two views looks sharper, as in conventional eye tests, the user adjusts the display to make the separate lines or dots come together and overlap, which corresponds to bringing the view into sharp focus. The underlying principle is similar to that used by new “adaptive optics” systems that have recently allowed ground-based telescopes to exceed the performance of the Hubble Space Telescope; these sometimes use the same kind of Shack-Hartmann sensors used in eye testing aberrometers.

The team will be field-testing the device in the Boston area this summer and will later test it in developing countries. The team already has applied for a patent on the system, named NETRA (Near-Eye Tool for Refractive Assessment), and team members won a prize this year in MIT’s annual IDEAS competition—a contest for inventions and business ideas that have a potential to make a significant impact in the developing world—and were finalists in the 2010 student-run MIT $100K Business Plan Competition.

Chika Ekeji, a student at the MIT Sloan School of Management who joined the team to help with commercialization of the system, says the group plans to launch production of the device as a for-profit company called PerfectSight, initially targeting parts of Africa and Asia. Ultimately, they also hope to produce a more advanced version that can incorporate its own higher-resolution display and be able to detect other conditions such as cataracts, which could be sold in the developed world as well. IT consultant Margaret McKenna is also assisting the team.

By using high-resolution LCD displays with this system, it is potentially not only much faster than today’s standard methods, but also “potentially more accurate,” Raskar says, although that hasn’t yet been demonstrated.