Israeli startup Umoove has attracted a lot of attention for its head- and eye-tracking technology after early reports that it might be powering a feature that does the same in Samsung’s Galaxy S IV. Now Bloomberg is saying the new flagship smartphone won’t launch with that tech on board. And while Umoove is keeping mum about any active deals it has with OEMs or other partners, the startup is being very vocal about its intent to make accurate head and eye tracking an everyday part of a user’s mobile experience.
The New York Times’ Brian Chen reported on Tuesday that Umoove’s platform will be open to everyone, with a new SDK private beta that launched this week to applicants. In an interview with TechCrunch, Umoove co-founder and CEO Moti Krispil said that they’ve already had hundreds of applicants, big and small, express their interest, and are planning a full public beta launch for April.
Krispil pulled back the curtain to give me a sneak peek at the tech behind Umoove’s platform. He explained how the startup hopes to improve drastically on tech that’s been around in one form or another for at least a decade to make it feasible to reproduce not only at scale, but with existing hardware and for a variety of budgets.
Umoove’s patented technology is designed to make the best possible use of the most basic raw data provided from any camera, including those that can only manage 320×240 (QVGA) and 10fps capture (below the capabilities of even many budget smartphones on the market today). It even works on a two-and-a-half-year-old iPod the company uses for testing, Krispil says. The platform then uses algorithms to fill in the gaps in whatever low-quality data it has captured through “a sophisticated use of real-time head/movement prediction components,” according to Krispil.
The key to Umoove’s ability to deliver consistently accurate results is that it is good at accounting for variances in environment, lighting conditions and user hand shake/movement. The technology is designed to provide a consistent experience, whether you’re in a brightly lit room or a darkened hall, and to work fluidly between them by adapting its processing when it detects color and brightness shifts. It uses an active stabilization technique to filter out natural body movements from an unstable camera in order to minimize false-positive motion detection, too.
Maybe most interesting of all, Umoove’s tech can work with partially obscured faces, meaning it can easily handle a user donning shades on a sunny walk. That means it can work even when a hand is partially obscuring the camera’s view of a user, enabling it to work better in tandem with touch tech. That, Krispil says, is one of the key reasons head and face tracking are needed over and above what Kinect and Leap are doing with full-body/gesture tracking. Especially on mobile, where the user is typically very near to the device, the other types of 3D controls aren’t really viable.
Another key element of Umoove’s platform is that it is optimized to work with lower-end processors, so that all devices can get in on the mix. The aim is to be inclusive as possible, Krispil says. Even if it does partner with some OEMs to integrate the tech at the system level, the hope is to become the provider of choice, so that small developers can also build it into their apps and have confidence that it will also work across platforms. CPU load on standard devices is at just between 1.5 to 5 percent, Umoove tells us, and on the Galaxy S3 takes less than 2 percent operating with all features, engines and algorithms turned on (some elements can be disabled to further save cycles).
“Nobody today is issuing less than a VGA in the last year, so even those smartphones that are $70 are great for us,” Krispil said. “If my daughter can play an arcade game on an iPod from two years ago, then there’s no reason anyone shouldn’t be able to benefit from our tech.”
Krispil closed by saying that Umoove can definitely work closer with hardware makers when and if appropriate to integrate hardware directly and make for an even better experience. (The company is already in some early discussions to do so, though he can’t reveal specifics.) But Umoove’s approach is two-fold, and also includes making sure the tech is applicable in a far-reaching, generic way for everyone to take advantage of.
For now, Umoove is gating access, and will be working with a select few during its closed beta. When it does open up to the general public, Krispil says it’ll have a volume-based, licensing fee structure, with enterprise options available for OEMs and other big clients.
Once again, the goal is to make sure everyone has access, regardless of their budget. It’s an ambitious goal, but judging by the reaction to my post from this past weekend, there’s a strong appetite out there for what it and other similar technologies can provide.
Umoove takes user experience to the next level by tracking and identifying gentle head movements and face gestures in real time on any mobile device using them to control applications & media. Their platform will be available soon to third parties who wish to add natural user interface to their application.