International Trade Today is a service of Warren Communications News.
Finger-Based Operation

Israel-Based Gesture Recognition Company Takes on Home Control With $199 System

Hoping to do what some technology companies haven’t successfully done, Israeli gesture recognition company EyeSight is bringing to market a plug-in device that controls connected devices by finger motions. The AC-powered device, which resembles a horizontal portable speaker, packs Bluetooth and Wi-Fi radios, an infrared code database for 150,000 devices and computer vision software and sensors, CEO Gideon Shmuel told us on a press tour in New York.

Sign up for a free preview to unlock the rest of this article

If your job depends on informed compliance, you need International Trade Today. Delivered every business day and available any time online, only International Trade Today helps you stay current on the increasingly complex international trade regulatory environment.

Calling the SingleCue a “new, smarter, fun way to control devices in the home,” Shmuel said the gesture-based device’s distinctiveness is combining IR control of legacy products -- TVs, set-top boxes and soundbars -- with control of contemporary digital devices such as media streamers and smart LED bulbs. Shmuel and Vice President-Product Management Tal Krzypow demonstrated finger-based control of an Apple TV media player through a Vizio TV along with the Philips Hue lighting system. Adding Z-Wave for more smart home integration is on the roadmap, Shmuel said.

Most remotes work in a space of “disparate devices,” said Krzypow. SingleCue is “not about the devices” but the activities that can be built around related devices, he said. In a typical home entertainment setting, the devices are supposed to work together, “but you have to operate them separately,” he said. With SingleCue, “we bind these devices into one experience and activity," he said. When one launches cable activity, "the TV turns on, the set-top box turns on, your soundbar … and the TV switches to the right input,” said Krzypow. “With one click, all these devices come to life with the right settings to work together.”

While Shmuel was able to whiz through control of the Apple TV, soundbar volume and lighting using minimal motion commands, our first effort ended in failure. The company boasts a 15 minute-to-master time with SingleCue. We didn’t allocate the full 15 minutes and don’t know if we would have mastered it in that time or not, but the device didn't register our finger puppet movements after several attempts. Microsoft Kinekt users, better trained in gesture control, are likely to have far better results than we did.

Gesture is the featured control method in SingleCue at launch, but the company plans to open the platform to voice and other sensing technologies, said Krzypow. The company expects a voice control feature sometime next year but Krzypow wouldn’t commit to a date. EyeSight is working on multi-modality control to fit interactions to context, said Shmuel. Changing channels by hitting “next, next, next” using gesture control would be cumbersome, he said, where simply speaking the word next repeatedly gets viewers there much faster.

There’s no single modality that can stand perfectly on its own,” said Krzypow. “It’s the combination of several that gives the great experience to the user because each has advantages and disadvantages.” They also often complement each other, he said. For SingleCue’s debut, “gestures are superior to voice if only because of the international aspect,” said Krzypow, citing the challenges of accents and ambient noise. Shmuel said voice control hasn’t taken off for TV operation due to complexity even with narrowly focused beamforming technology. It’s not practical for viewers to speak into a mic on a remote to issue commands, said Shmuel. “You have to look for your remote control and speak to your remote control, and I think it defies the purpose,” he said. “The point is to not have to look for the remote control.”

SingleCue translates what machines see from a finger -- pixels -- into actions and features, said Shmuel. When a user raises a finger, a sensor reads the finger movement, and the embedded processor runs computer vision algorithm software that scans the space, said Shmuel. The robust software is constantly "looking for suspects” such as a finger tip, he said. When a finger is raised in range of the sensor, SingleCue picks up the finger and detects "changes in the object," he said.

While gen-one SingleCue targets the home entertainment space, EyeSight has sights set on broader automation and the automotive space, Shmuel said. The company works with AMD, Lenovo, Philips and Toshiba on embedded solutions for gesture control, he said. When it decided to come out with a new product category, it was easier for the company to build the device itself using off-the-shelf components and a powerful Freescale processor than to farm out the box to an OEM, said Shmuel. In cars, the company's software could be used to sense a nodding head or even the emotion of anger. Years ahead, that could trigger the car to insert space between the vehicle and the next one ahead, said Shmuel.

SingleCue has a Thursday ship date from Amazon for $199, and the company is planning an IR repeater for Q1. Range of the SingleCue is about 15 feet, so the repeater will extend remote control of IR devices farther and to components stored behind doors and not within line of sight. Asked about the competition, Shmuel said there's nothing like SingleCue in the market, although its functionality is similar to that of a universal remote or smart hub that connects to an app. Companies that make those devices are talking to EyeSight about incorporating gesture control into their products. Shmuel cited the new “multi-modal” control world where consumers will interact with devices via various means: gesture, voice and touch, for instance. Companies who want to add gesture will turn to EyeSight, for one, instead of developing their own technology, he said.

EyeSight’s mission is “to bring the natural user interface everywhere,” said Krzypow. SingleCue can “control the devices you already have at home even if they have no camera or software preloaded,” he said. That means integrating with a lot of other companies, he said. He envisioned a scenario where a user could control the TV from the kitchen using a device such as Amazon’s voice-enabled Echo to communicate with SingleCue. “That’s the experience we’d like to enable ultimately,” said Krzypow.

The future of home entertainment control is embedded technology, and EyeSight is doing it, said Shmuel. The company has been embedding its algorithms in chipsets and is now stuffing them into its own device, but it also wants to have OEM agreements with service providers. “We’re already in discussions with big cable companies to incorporate our functionality into their set-top boxes,” he said. Instead of having the user interface on the cable box, it could be on the TV using EyeSight technology, Shmuel said.

On whether the company can make money off the SingleCue box, Shmuel said, “That’s the plan.” He contrasted the business model with that of TVs where “it’s very hard to make money.” TV companies that are successful either make panels like Samsung or operate “very lean” like Vizio, he said. Similarly, SingleCue was designed with off-the-shelf components, “the right components to produce the right experience,” said Krzypow. OEMs aren’t willing to “go all the way” and truly adapt an interface for a new interaction, he said, so EyeSight went on its own. “A new interaction needs its own interface," Krzypow said. “Because of that, with SingleCue, for first time we could build an experience designed for touch-free gesture.”