A Start-Up Has Survived to Beam Up Your Gestures

The founding team of the small Silicon Valley chip company set out 11 years ago to give any electronic device with a tiny digital camera the ability to view the world around it in three dimensions. Such that mimics the human eye holds the promise of altering the ways in which humans interact with the computers that fill their lives. But later this year, when consumers begin to find computers, TVs and gaming systems in stores that they can control through hand and body , the company may finally escape its many near-death experiences.

“The 3-D cameras used to cost $5,000,” said Michel Tombroff, the chief executive of Softkinetic, which makes software for the 3-D gesture recognition systems. “That price has come down to $50 or less for the bill of materials.”

Advocates of this technology hail it as one of the most significant changes to the way people interact with their devices. Some even expect to it to vastly overshadow the touch-screen technology that has recently captured attention.

But the technology has proved so difficult to create that Canesta, based here, has spent its entire existence selling far more promises than actual products. In 2002, James Spare, then the head of marketing at Canesta, began showing off a virtual keyboard to the public. Using Canesta’s technology, hand-held devices like the Palm Pilot and cellphones could project a keyboard, in flickering red light, onto a desk or table. No one would have to lug a keyboard.

Canesta’s virtual keyboard garnered plenty of attention, but it did not work all that well. People were used to putting pressure on keys, while Canesta’s system required them to think more about the position of their fingers in space. Human anatomy proved troubling as well.

“Fingernails were a problem,” said Cyrus Bamji, the chief technology officer at Canesta. “We could not see fingernails because they would disperse the light sent out by our laser. We joked about selling a line of special fingernail polish.”

Canesta’s technology had some early industrial uses in the automotive and robotics fields. Car companies could create sensors that would help vehicles see obstacles around them, and industrial robots could be equipped with sensors for checking product quality.

Still, Canesta’s executives thought that a bigger payoff would come by burrowing into the consumer electronics market. But to get into everyday gadgets, Canesta would have to create a chip cheap enough and tiny enough to be paired with minuscule digital cameras. The chip module would send out infrared light and then calculate how long that light took to return to the module.

The company’s engineers knew that building such a chip would require herculean feats of creativity and years of toil. For one, they had to figure out a way to coax the chip into seeing only Canesta’s brand of infrared light while ignoring the immense amount of infrared energy from the sun.

“I remember a meeting in Japan where an executive opened the window ahead of our demonstration,” Mr. Bamji said. “There was a lake outside, and we got all this reflection back from the sun. That was a big, embarrassing problem.”

After such setbacks, Canesta needed more money. That meant persuading investors to take a big risk on the unknown.

“We had 140 meetings with venture capitalists,” said Mr. Spare, now the chief executive. “It has been a challenge to find investors that are focused on the big opportunity.”

Investors in Silicon Valley have largely given Canesta the cold shoulder. Long-term investments in chip companies have become increasingly rare here.

Over the years, Canesta has secured about $60 million. Investors like the Carlyle Group and Venrock have stuck with it, while other investors from Taiwan and Korea have joined in.

Notably, Quanta, a Taiwanese company that manufactures laptops for the top brands, has put money into Canesta. It expects to sell laptops this year or next with the modules. Honda has invested as well, and there is talk of using Canesta’s chips for sensors on the outside of cars and internal sensors that can see whether an adult or a child is in a seat and adjust the way an air bag behaves accordingly.

Over the next year, PCs and TVs will arrive with the company’s chip inside. People will soon be able to turn on their TVs simply by raising their hand in the air or flip through photos on their computers with a wag of a finger.

Canesta’s gaming smarts were on display at the South by Southwest conference in Austin, Tex. A company called Frog Design held a party where people could play classic arcade games like Breakout and Galaga with their bodies on a 12-foot-by-9-foot screen, using Canesta’s technology.

In the coming years, the expectation is that just about every digital camera in consumer devices like PCs and cellphones will be replaced by a 3-D camera capable of recognizing gestures. Microsoft has sparked tremendous interest in this gesture recognition technology through a gaming system code-named Project Natal. This is a device that Microsoft will sell later this year for its Xbox 360 consoles, allowing people to do away with controllers altogether and use their bodies to play games.

Canesta hopes to piggyback on the exploding interest in the technology and pop low-cost chips into every device it can.

“We will make a little profit on every one of those chips,” Mr. Spare said. “Years ago, there may have been a light at the end of the tunnel, but it was hard to see and blinking at best. Now that light is bright.”

Leave a Reply

Your email address will not be published. Required fields are marked *