Step right up, ladies and gentlemen, and behold…the eighth wonder of the world. No, not King Kong, I’m afraid. But how about a computer that can read and interpret human emotions and mental states?

That would be from Affectiva, a Boston-area startup co-founded by Roz Picard, a 20-year veteran of the MIT Media Lab. Picard (see photo, right) is the founder and director of the Media Lab’s Affective Computing research group, and she has done extensive work in computer vision, machine learning, and human-computer interfaces, with applications in autism communication, health and wellness, education, marketing, advertising, and other areas.

Picard is speaking at this Thursday’s Xconomy “6×6” event (Six Cities, Six Big Tech Ideas) in Boston. In advance of her appearance there, I sat down with Picard at Affectiva’s offices this week to get a demo of the company’s technology and to talk a little about the future of emotional and gestural interfaces.

One of the demos involved a computer tracking my facial expressions via webcam (and also my heart rate via blood flow to my face) while I watched a series of TV commercials. Based on indicators like raised eyebrows, smiles, or a furrowed brow (see example, below), the software tried to figure out how engaged, interested, amused, or disturbed I was during the course of each ad. It’s hard not to be self-conscious during all this, but I’m pretty sure the computer concluded: this guy hates all commercials. (And since I haven’t smiled since 1995, we’ll enlist my fellow editor Erin Kutz for the live demo on Thursday.)

Affectiva also will be rolling out a new product on Thursday—one that has applications in finance and healthcare, among other industries—but I’ll let Picard speak for that when the time comes.

Meanwhile, I asked Picard whether the field of affective computing would continue to advance incrementally (like speech recognition, say) or whether it would undergo a breakthrough of some kind. “I think it’s going to make some leaps,” she said. “There’s going to be a lot more happening by indirect measurement—nonverbal [cues] that people don’t really think machines can do. That’s going to really progress.”

To me, Picard’s work exemplifies what truly big ideas are about—for the first 20 years or so, they might be more interesting scientifically than commercially. But once the technology and marketplace gets to a certain point, a viable business can be built around them, even as the science continues to advance. And then who knows what will happen?

We’re looking forward to a fantastic 6×6 program and some great networking this Thursday (you can register here). Hope to see you there.

Comments (3) | Reprints | Share:          

Qualcomm (NASDAQ: QCOM), the San Diego wireless technologies giant, said today it has acquired certain assets related to gesture recognition technology from Sunnyvale, CA-based GestureTek—a move that points to new capabilities at its new Atheros unit.

Financial terms of the deal were not disclosed. GestureTek, founded in 1986, specializes in machine vision technology that enables people to use hand and body motions to dynamically control computer-based information displayed on a screen or camera-enabled device. GestureTek says there is no need for users to wear, hold, or touch anything special for its proprietary technology to work.

In a statement, Qualcomm group president Steve Mollenkopf, says, “Applications processors are enabling a range of new ways for consumers to interface with their home entertainment and mobile devices. Our acquisition of key technology and assets from GestureTek will strengthen Qualcomm’s smartphone product portfolio and enable our customers to launch products with new and compelling user experiences.”

The acquisition appears tailor-made for Qualcomm Atheros, the business unit formed in May, after closing its acquisition of  San Jose, CA-based Wi-Fi chipmaker Atheros Communications for $3.1 billion. Qualcomm has explained the combined companies will advance development of products for “the connected home,” such as wireless technologies for displaying information from smartphones on high-definition TV screens and larger monitors.

Qualcomm says the GestureTek technology will be integrated into Qualcomm’s current and next-generation Snapdragon processors, providing possibilities for new types of user interfaces from Qualcomm customers making smartphones, tablets, and video game consoles. In the latter market, Qualcomm could be challenging established motion-recognition technologies developed for the Nintendo Wii, Playstation and Xbox360 video game console.

GestureTek has retained other assets to continue its its gesture-controlled public display and digital signage business. The company’s business has been focused on displays for kiosks, the healthcare industry, and consumer electronics.

Comments (3) | Reprints | Share:          

Waltham, MA-based Affectiva, an MIT Media Lab spinout, said today it has raised $5.7 million in Series B financing led by Kantar (WPP) and Myrian Capital. The company says it will use the money to develop and market its various software products for emotion recognition. Its technology includes computer vision and machine learning algorithms that estimate emotional states based on facial expressions (via webcam), as well as biometric sensors and software that track things like excitement, engagement, and anxiety. Affectiva was founded in 2009 by MIT scientists Roz Picard and Rana el Kaliouby; the company is led by CEO Dave Berman.

Comments (1) | Reprints | Share:          

San Mateo, CA-based VideoSurf said today that it has raised $16 million from Pitango Venture Capital and strategic investors, bringing its total financing to $28 million. Using a patended computer vision algorithms to identify Web video content, Videosurf has built a search index that it says is used by 12 million visitors per month. VideoSurf CEO and co-founder Lior Delgo said the new cash “gives us the resources to scale the company as we continue to hire the best talent and invest in our technology.”

Comments | Reprints | Share:          

The world needs robots, hopefully more than Mars needs moms.

The first part of that was Rodney Brooks’s message in his talk at the annual meeting of the National Venture Capital Association in Boston yesterday. Brooks, of course, has a strong interest in seeing the mass adoption of robots. He’s the co-founder of iRobot (NASDAQ: IRBT) and Heartland Robotics, the Boston startup that aims to revolutionize U.S. manufacturing. Brooks retired last summer from an illustrious 29-year academic career, mostly at MIT, to focus on Heartland, which raised a $20 million Series B round last fall.

Here are a few takeaways from the Brooks talk, which was delivered to a room of venture capitalists:

Brooks started by making the economic case for mainstream robots. The populations of Europe, Japan, and the U.S. are getting older, he showed—the ratio of elderly to young people is increasing. That means “older people are competing for the physical services of younger people, so the price goes up too,” he said. “Physical stuff that needs to be done for people will get much more expensive.” And robots are a way to help do that stuff more efficiently, he said—think robotic wheelchairs or simple machines to lift and carry groceries and heavy items.

Brooks then traced the geographic path of low-cost manufacturing since World War II. He showed that the U.S. has outsourced cheap industrial labor to a series of countries—first Japan; then Korea, Taiwan, and China; and then Vietnam, Malaysia, Thailand, and Indonesia. Next up: Burma, Bangladesh, and perhaps African nations. “I don’t think this is sustainable in the long term,” Brooks said. “Everyone comes up in this flat world.”

He didn’t give an overt commercial for Heartland Robotics—the company is still pretty guarded about its technology—-but the idea (as I understand it) is that inexpensive helper robots in factories and workplaces could increase U.S. competitiveness, and eventually lead to what Brooks calls the “third wave of the industrial revolution.”

I gleaned a few more hints about what Heartland’s robots might be able to do. Judging from Brooks’s research over the years in computer vision, object recognition, dexterous manipulation, and interactive robots, I’m guessing they’ll be aware of people and objects around them and what exactly they’re looking at. They’ll be able to pick up and manipulate simple items, all while being safe for people to be in close quarters with. Most of all, they’ll be cheap and easy to use—a new breed of … Next Page »

Comments | Reprints | Share:          

San Francisco-based Affine Systems, which uses computer vision technology to identify the content of online videos and help advertisers match ads with related content, said yesterday that it has raised $5 million in Series B funding. New investor Crosslink Capital led the round, which was joined by existing backer Highland Capital Partners. The company has raised $8 million all told. “We plan to use Crosslink Capital and Highland’s investment to scale our operations and bring advertisers the tools they need to be confident that their ads will be shown in connection with the most relevant and appropriate online videos,” Affine CEO Mike Sullivan said in a statement.

Comments | Reprints | Share:          

It’s a staple of every film depiction of killer androids since Terminator: the moment when the audience watches through the robot’s eyes as it scans a human face, compares the person to a photo stored in its memory, and targets its unlucky victim for elimination.

That’s computer vision in action—but it’s actually one of the easiest examples, from a computational point of view. It’s a simple case of testing whether an acquired image matches a stored one. What if the android doesn’t know whether its target is a human or an animal or a rock, and it has to compare everything it sees against the whole universe of digital images? That’s the more general problem in computer vision, and it’s very, very hard.

But just as we saw with the case of statistical machine translation in Part 2 of this series, real computer science is catching up with, and in some cases outpacing, science fiction. And here, again, Google’s software engineers are helping push to the boundaries of what’s possible. Google made its name helping people find textual data on the Web, and it makes nearly all of its money selling text-based ads. But the company also has a deep interest in programming machines to comprehend the visual world—not so that they can terminate people more easily (not until Skynet takes over, anyway) but so that they can supply us with more information about all the unidentified or under-described objects we come across in our daily lives.

I’ve already described how Google’s speech recognition tools help you initiate searches by speaking to your smartphone rather than pecking away at its tiny keyboard. With Google Goggles, a visual search tool that debuted on Android mobile phones in December 2009 and on the Apple iPhone in October 2010, your phone’s built-in camera becomes the input channel, and the images you capture become the search queries. For limited categories of things—bar codes, text on signs or restaurant menus, book covers, famous paintings, wine labels, company logos—Goggles already works extremely well. And Google’s computer vision team is training its software to recognize many more types of things. In the near future, according to Hartmut Neven, the company’s technical lead manager for image recognition, Goggles might be able to tell a maple leaf from an oak leaf, or look at a chess board and suggest your next move.

Goggles is the most experimental, and the most audacious, of the technologies that Google CEO Eric Schmidt described in a recent speech in Berlin as the harbingers of an age of “augmented humanity.” Even more than the company’s speech recognition or machine translation tools, the software that Neven’s team is building—which is naturally tailored for smartphones and other sensor-laden mobile platforms—points toward a future where Google may be at hand to mediate nearly every instance of human curiosity.

“It is indeed not many years out where you can have this little bird looking over your shoulder, interpreting the scenes that you are seeing and pretty much for every piece in the scene—art, buildings, the people around you,” Neven told me in an interview late last year. “You can see that we will soon approach the point where the artificial system knows much more about what you are looking at than you know yourself.”

Going Universal

Neven, like most of the polymaths at Google, started out studying subjects completely unrelated to search. In his case, it was classical physics, followed by a stint in theoretical neurobiology, where he applied methods from statistical physics to understanding how the brain makes sense of information from the nervous system.

“One of most fascinating objects of study in nature is the human brain, understanding how we learn, how we perceive,” Neven says. “Conscious experience is one of the big riddles in science. I am less and less optimistic that we will ever solve them—they’re probably not even … Next Page »

Comments (2) | Reprints | Share:          

A San Francisco startup called Stipple has raised $2 million from big-name investors to help make images on the Web more interactive. Specifically, the company wants to help publishers who puts lots of photos online to turn them into gateways to various kinds of interactions, including online purchases.

When you think about it, images on the Internet are weirdly static. The Web is dripping with them, of course, but they lack the interactivity that we expect these days from audio, text, and video. Stipple’s technology lets anyone with his own website layer interactive labels atop each image, including captions and what it calls “dots” offering details and links about specific people, places, or things. The information is invisible until a Web user mouses over the image. (Check out the images on Stipple’s home page for a taste of how this works.)

“Generally speaking you’re doing the same thing on the Web with images today that you were doing 10 years ago,” says Stipple founder and CEO Rey Flemings. “They are the forgotten media type on the Web. There hasn’t been a tremendous commercial focus on making it super-easy for people to discover what’s in photos. That was the initial idea of Stipple.”

That idea resonates with customers who’ve been using Stipple’s technology since the company’s August 2010 launch, including blog hosting company Six Apart, the E.W. Scripps newspaper chain, and record labels JIVE Label Group and Atlantic Records. And it appeals to the startup’s eclectic group of investors—it’s not too often you see singer/actor Justin Timberlake and staid Silicon Valley venture firm Kleiner Perkins Caufield & Byers backing the same company.

Other contributors to the $2 million round include Mike Maples’ Floodgate Fund, Parkview Ventures, Quest Venture Partners, Global Brain Corporation, and individual investors John Ferber, Rick Marini, Eghosa Omoigui, and Naval Ravikant. In a press release announcing the investment today, Maples offer this buzzword-ridden yet succinct summary of the startup: “By marrying the ubiquity and engagement of photos with a business model and user experience that harnesses the power of community and network effects, Stipple has created the underpinnings of a company with disruptive potential in a massive market.”

Massive indeed: there are well over a trillion images on the Web, yet the only data available about most of them is their file names—and that’s usually something cryptic like “IMG_7149.jpg”. (Some images also come with metadata tags using the exchangeable image file format, or EXIF, but this usually only covers technical details like the date and time the picture was taken, or the settings of the camera used.)

Companies like Munjal Shah’s Riya have tackled the problem of making images more discoverable using computer vision techniques. Such algorithms are getting better at recognizing faces and certain kinds of objects, but aren’t yet able to … Next Page »

Comments | Reprints | Share:          

None of this would have happened 10 years ago. Where to begin?

Last month, I walked into a room of about a dozen robotics experts and technology startup investors. It was one of the sessions at the MassTLC Innovation “unConference” in Boston. The discussion centered around how to build a successful robotics company. But it was some of the newer context around this question that turned the session into a watershed moment I won’t soon forget.

Two main takeaways: First, robotics companies around Boston have come a very long way since 2000, when I was a postdoc in a robotics group at the MIT Artificial Intelligence Lab. That sort of academic research is still going strong, but the bigger story of the past decade has been the business success of robotic vacuum cleaners, bomb-disposal units, and surveillance drones, and how that has helped pave the way for a new generation of companies.

My second takeaway is that the business community thinks there is a new threat to Boston’s competitive position in robotics—and its name is Google. I’m not usually one to fan the flames of Boston vs. Silicon Valley arguments, but in this case the discussion hits close to home, so I wanted to see if there’s much truth to it.

The Boston area, of course, is home to numerous robotics companies—iRobot (NASDAQ: IRBT), Boston Dynamics, Harvest Automation, Heartland Robotics, Kiva Systems, iWalk, and CyPhy Works, just to name a few. Many of those companies were represented in the unConference session, along with investors from CommonAngels, Founder Collective, and General Catalyst. Historically the region has had lots of expertise, both in universities and industry, in key technologies underlying robotics such as sensors, actuators, control algorithms, artificial intelligence, computer vision, and data storage.

Yet 10 years ago, most early-stage investors (angels and VCs) wouldn’t think of touching a robotics startup. The development costs and business risks were too high, and the technology infrastructure—onboard processing power, wireless communication, programmable chips, sensors, algorithms—wasn’t quite ready for prime time. Now things have changed, certainly in investors’ minds, … Next Page »

Comments | Reprints | Share:          

A couple of startup companies set a world aviation record last night.

But they were pretty low-key about it. As I walked into the Future of Flight Aviation Center in Mukilteo, WA, a half hour north of Seattle, I saw little activity. It was after hours, and the hangar-like building was nearly deserted except for the futuristic planes suspended from the ceiling—Burt Rutan’s “Quickie” and a Beechcraft Starship—and part of a Boeing 787 Dreamliner fuselage on the display floor. It was a bit like “Star Wars” meets “Night at the Museum.”

Tom Nugent, the co-founder and president of Kent, WA-based LaserMotive, greeted me and said they were almost ready for showtime. A small team of engineers divided its attention between the back of a command truck and the adjacent trailer that held the laser optics equipment that would make the show possible. Two German guys who hadn’t slept in days (and were still on Munich time) were sprawled out on deck chairs in front of computer monitors like they were playing a video game. One held a remote controller that he used to guide a “quadrocopter”—a small, 1-kilogram, square-shaped flying contraption with blinking lights and four spinning rotors—made by their company, Ascending Technologies.

Jan Stumpf and Michael Achtelik, the co-CEOs of Ascending Technologies, partnered with LaserMotive to perform this feat last night. The goal: to use a laser to power an aircraft in continuous flight for about 12 hours (far longer than its battery would last without recharging, which is only about five minutes). That would be a world record, by a long shot, for the longest free flight of an electric vehicle.

Indeed, this demonstration is a big deal for the future of electric planes, said Barry Smith, the executive director of the Future of Flight facility. Imagine putting a laser on top of every cellular tower, he said, so that certain types of unmanned aerial vehicles (UAVs) would never need to land to recharge or refuel. That could potentially revolutionize communications, surveillance, and security and defense applications. Longer term, it could even impact the long-held dream of powering manned aircraft with electricity instead of jet fuel—though that is very far off.

For now, Nugent says, “The significance is we’re going to show this quadrocopter, and any aerial vehicle [of this size], will be able to fly effectively forever. It’s no longer limited by battery capacity.”

LaserMotive has done smaller flight tests before, but not on a free-flying vehicle like this. The company is best known for winning the $900,000 NASA Power Beaming Challenge last year, in one of the levels of the “Space Elevator Games.” That involved using a laser to power a climbing robot up a cable to a certain height (1 kilometer) at a certain speed (about 9 mph). But lately the company has been targeting UAVs as a big commercial application of its wireless power technology. (The next level of the NASA challenge, which was supposed to happen later this year, is still up in the air, so to speak.)

“Goggles on!” someone shouted, and we all complied. That meant the infrared laser, which puts out about 200 watts of light power, was switching on. The beam was directed using a series of mirrors and optics and shot out the top of the trailer. You couldn’t see it with the naked eye except for a reddish halo on the 50-foot ceiling. At the same time, the quadrocopter lifted off (under its own battery power), guided by Stumpf, and floated up to meet the beam, about 30 feet off the ground (see left).

“Not centered,” Nugent said. Then the computer vision system of LaserMotive’s setup kicked in. Software and cameras aligned with the path of the laser beam tracked the vehicle’s position, and positioned the beam so it hit the photovoltaic cells on the underside of the craft; those solar cells transformed the laser’s energy into electricity to continuously charge the quadrocopter’s battery.

With that, all human corrections fell away, and it was just a drone hovering eerily in space, rotors humming quietly. It swayed a few feet from side to side, and the laser tracked it. It was about 7:40 pm.

This is the boring part, Nugent said. And boring is good. Exciting is bad. For the next 12 hours, if all went well, nothing more would happen. The craft would stay up all night (as would the crew), … Next Page »

Comments (2) | Reprints | Share:          

If you’re in the Faneuil Hall area of Boston today, look out for the trike. That would be the Google Street View Trike, a pedicab-like contraption mounted with a camera and computer equipment. It’s meant to capture 360-degree, street-level images of places where Google’s fleet of Street View cars can’t go—pedestrian malls, university campuses, parks, hiking trails, and so forth.

As my colleague Wade reported last November, Faneuil Hall was a finalist for the trike treatment, competing against Chicago’s Navy Pier and San Francisco’s Pier 39 in the pedestrian mall category (via online voting). Well, the cradle of liberty won out, and today, Google is showing up to take the pictures. They should be viewable in Google Maps sometime in the coming months, the company says. (You can read more about the nuts and bolts of the trike project in Wade’s interview with senior mechanical engineer Dan Ratner.)

Of course, it’s all great marketing for Google. But it’s also a really interesting step in the evolution of digital imaging technology, and how consumers can experience the richer details of the real world while online. And it could eventually tie into Google’s local search business.

So I decided to do a deeper dive into Google Street View, which is an ongoing project in all 50 U.S. states and dozens of countries around the world. Yesterday I spoke with the mastermind of the Street View project, engineering director Luc Vincent, who’s based in Silicon Valley.

Vincent is a renowned expert in image processing and computer vision. A native of France, he was a postdoctoral fellow at Harvard University’s Robotics Laboratory in the early 1990s before developing his career at Xerox Imaging Systems in Peabody, MA. From there, he moved out west and spent time at ScanSoft, Xerox PARC, and LizardTech, before joining Google in 2004, originally to work on Google book search.

We talked about the genesis and history of Google Street View, as well as the future of geo-search and imaging—its significance to the company, how far Google wants to go (hint: everywhere), and how it deals with privacy issues. Here are some edited highlights from our chat.

Xconomy: So how did Street View originally come about?

Luc Vincent: I worked mostly on books for the first two years. But the same day I joined Google, I was put in meetings about collaborations with Stanford to do research [on gathering images]. I turned this into a Google “20 percent” project, and found some more people to work on it. Pretty soon I was herding cats.

Larry Page himself was interested in this when I joined Google. We were not really motivated by money originally—just building compelling services. We focused on the scale. We were really willing to spend money and engineering [resources] because we thought it would be useful to people. At Google, we start small, we show demos, and we iterate.

X: What was the biggest challenge in getting the project going?

LV: There were tons of challenges. Early on, we had no funding per se. We got people to help with time and equipment. When we started with Stanford, we were working with the DARPA Grand Challenge [robotic car] team. We used [one of their cars, with a driver] to collect our first test imagery. These cars were too fancy and automated, so we drained the battery multiple times. If you drain the battery, the A/C stopped working. It was summer and we were trying to collect data, and we had … Next Page »

Comments (3) | Reprints | Share:          

Like.com, a four-year-old startup in San Mateo, CA, that created a “visual shopping” site underpinned by sophisticated computer vision and machine learning algorithms, has been acquired by Google, according to an announcement on Like’s website. “We see joining Google as a way to supersize our vision and supercharge our passion,” CEO and co-founder Munjal Shah wrote in the announcement. Terms of the deal were not revealed. The 24-employee company had raised at least $54 million in venture backing from Bay Partners, BlueRun Ventures, Crosslink Capital, First Round Capital, Leapfrog Ventures, and Menlo Ventures.

Comments | Reprints | Share:          

Famed robotics expert Rodney Brooks, the former director of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), is retiring from academic duties at MIT as of this week. The co-founder and former chief technical officer of Bedford, MA-based iRobot will be focusing full-time on his newest company, Heartland Robotics, based in Cambridge, MA. Brooks’s new role as professor emeritus at MIT will be effective this Thursday, July 1. He is 55.

There is some question as to whether Brooks will be the youngest professor emeritus in the history of his department (electrical engineering and computer science). Reached by e-mail, Brooks confirms he is the youngest of the current crop of MIT professors who are retiring this year—at least those on a list announced at a faculty meeting last month. And current regulations at MIT say a professor must be at least 55 to achieve emeritus status, so he is probably among the youngest, at least in recent history. “It may be that come this Thursday I will be the current youngest emeritus professor at MIT, but I can’t say that for sure,” Brooks says.

Brooks is widely known for his scientific contributions to computer vision, mobile robots, humanoid robots, artificial intelligence, and artificial life. He did postdoctoral research at Carnegie Mellon University and MIT before becoming a professor at Stanford University (where he had done his PhD in computer science) and then MIT in 1984, where he has stayed as a full-time faculty member until now. Brooks was director of the MIT AI Lab before it merged with the Laboratory for Computer Science to form CSAIL in 2003. He stepped down as director of CSAIL in June 2007 to dive deeper into science and, as it turns out, to try to create a new industry.

“I spent my career at MIT developing new ideas for how to make robots intelligent. Through iRobot I have been very successful at getting robots out in the world based on those ideas. In terms of raw numbers the majority of robots that are sold today are behavior-based, and follow the ideas I worked on in the eighties,” Brooks says. “Now I am trying to develop a new class of robots for deployment in the real world that are based on the work that my students and I did during the nineties and later.”

In September 2008, Brooks announced he was leaving iRobot (NASDAQ: IRBT), the consumer and defense robotics firm he started in 1990, to build a new company. Heartland Robotics, which is backed by Charles River Ventures and Amazon.com founder Jeff Bezos, is looking to bring intelligent, dexterous robots to industrial workplaces in order to boost productivity and revitalize U.S. manufacturing. It’s still very early, of course. But given Brooks’s recent advances in robots that can manipulate … Next Page »

Comments (2) | Reprints | Share:          

[Updated 3:30 pm, 4/23/10. See below] Zoran Popovic, a computer scientist at the University of Washington, is constructing 3-D virtual recreations of real life. First step, the UW campus; next, the whole world.

Building on a previous program called Photo Tourism that pieces together photos culled from Flickr into virtual 3-D models, PhotoCity is a “capture the flag”-esque game that recreates sections of campuses or city blocks, or, eventually, entire cities from user-generated photos (see reconstructed image, above right).

Photo Tourism was created by UW professor Steve Seitz, Microsoft researcher Richard Szeliski, and former UW graduate student Noah Snavely (now an assistant professor at Cornell University), and in 2006 was licensed to Microsoft. But that program was limited, Popovic said: While they could easily build a 3-D version of the Coliseum based on tourist photos deposited in Flickr, they were never going to get a usable model of the building next to the Coliseum, for example. [The list of people involved with Photo Tourism and PhotoCity has been corrected and updated---Eds.]

Popovic is interested in what he calls “serious games,” computer games that rely on volunteer user input to solve problems too difficult for computational power alone. Usually, players are motivated to help solve a problem by the lure of some kind of online competition. But recreating 3-D versions of city blocks or whole cities adds another layer of challenge—the game has to be compelling enough not only to get people to sit down and play it on their computers, but to take extra digital pictures for the sole purpose of expanding the 3-D world. “It occurred to me that one big problem is that you have to step away from the computer,” Popovic said. “So we wanted to know, is it possible to form a game framework where people go out and do stuff in the real world, instead of just sitting in front of their computers?”

To motivate people to take more pictures for the models, the group (including UW graduate students Kathleen Tuite and Dun-Yu Hsiao) created the PhotoCity game, which can be played on a computer or iPhone. A Google Maps image shows the playing field (so far, game locations include the UW and Cornell campuses, and neighborhoods in Seattle, New York, Boston, Chicago, Santa Barbara, Portland, Washington DC, and Moscow), and “flags” pop up at various points on the rough facade. Players then win these flags by taking enough pictures of that site—say, the northwest corner of UW’s Odegaard library—and then uploading them to the PhotoCity site. Win enough flags and you’ll eventually be the “owner” of the whole building. Other players can steal flags away by taking even more pictures.

The researchers launched the game at the end of March by announcing a competition between UW and Cornell, to see which school’s team could get the most photos of their campus by April 20, but soon other locations were added. Any user can “seed” a new playing field. The cross-country competition is now over, but results are still being analyzed, Popovic said, and the teams are planning a second round to start up April 30.

Popovic (left) is happy with how the game has progressed, but his group’s eventual goal is loftier: to recreate entire cities and, eventually, the world in 3-D from digital pictures. “If you can build a 3-D replica of your world, you can all of a sudden have all different kinds of games in that world,” Popovic said. “The idea is to make a game where the actual world itself is built, and people can actually become avatars and walk around and do … Next Page »

Comments (1) | Reprints | Share:          

This week, I’ve learned a lot about game-changing ideas and how to think about making them work. Like anything meaningful, some of the lessons will take more time and effort to sink in. But here are five lessons to take away from our Xconomy Forum (“What’s Your Breakthrough Idea”) held at the University of Washington on Monday:

1. Not everything is, or should be, a breakthrough idea. Nick Hanauer of Second Avenue Partners framed the whole discussion by pointing out that entrepreneurs have different motivations: some might want to make a lot of money whether or not they change the world; others may want to change the world whether or not they make much money. Either approach is perfectly valid; just be true to yourself.

2. Venture capital alone won’t sustain breakthrough ideas. VCs don’t fund new ideas or the invention process, said Nathan Myhrvold of Intellectual Ventures. Instead, they fund “zillions of ‘me-too’ ideas,” he said. Which is why Myhrvold is trying to create a new “invention capital” marketplace—and why his company has awarded $315 million to individual inventors in the U.S. and has deals with more than 100 universities to support the invention process.

Meanwhile, Hanauer told me last week that most VCs don’t take risks anymore because the VC business model rewards those who can simply avoid a major screw-up. “The business model is toxic to risk-taking, because it’s so unbelievably profitable for the partners just if it doesn’t fail,” he said. (Of course, VCs will tell you that their model isn’t broken—because it isn’t.)

3. The proper mindset of breakthrough-idea thinking is to be “narrowly insane but not a total whack job,” as Myhrvold put it. By narrowly insane, he meant that inventors need to be crazy enough to think they can do something unprecedented, without being delusional. Put a different way, it’s helpful to be a “high-functioning contrarian,” as Hanauer says Jeff Bezos has described him. (“A low-functioning contrarian means you’re in prison,” Hanauer adds.) In other words, try to see the world differently, and imagine what would happen if things were arranged in other ways. Amazon.com, for example, delivered more than 10 times the selection of a brick-and-mortar store at a cost savings of more than 25 percent. That’s the kind of thinking that can transform an entire industry—in this case, those who sell books (for starters).

4. Part of being a good entrepreneur means not listening to your critics, or the entrenched interests. As Lee Hood from the Institute for Systems Biology and Integrated Diagnostics put it succinctly (I’m paraphrasing), people have seriously doubted him six or seven times in the past—and he’s been right every time. (He didn’t say how long it took to be proven right.) But the basic message was that if you want to change the world, you will meet with resistance—the people and companies in power don’t want things to change—but don’t let that deter you. Hanauer stressed the importance of … Next Page »

Comments (2) | Reprints | Share:          

There’s no crying in baseball—or in innovation. Yet I’ve been hearing a fair bit of lamenting around town about how Seattle startups or venture capitalists are not making the top-whatever lists (Wall Street Journal, Technology Review, you name it) of most innovative companies or investors around the country.

Who cares? The city that brought the world Microsoft, Amazon, Boeing, McCaw Cellular, RealNetworks, Starbucks, and any number of other game-changing companies, should not be too worried. It should, however, be motivated. And it is.

That’s why we’re gathering an elite group of technology and life sciences leaders to talk about how to spot the biggest, most promising ideas in their fields, and how these ideas could really change the world, at our half-day Xconomy Forum, “What’s Your Breakthrough Idea?” on March 29 at the University of Washington. (Tickets are going fast, but there are still a few left at our special saver rate—you can register here.)

This is about thinking big as a community. But it’s also about the realities around making the biggest ideas work—building the right teams for the right companies, understanding who the customers are, handling competition, and so forth. And although the focus will be mostly on Seattle-area companies and leaders, the impact that we’re talking about here is decidedly global.

Here’s a little more about the program and speakers for March 29 (topics are subject to change):

—Nick Hanauer of Second Avenue Partners will open with a keynote about the mental calculus behind identifying a true breakthrough idea, versus a really, really good idea that probably won’t get traction. Nick doesn’t think of himself as a technology guy, yet some of his blockbuster investments happen to be in tech—Amazon, aQuantive, Insitu. (There’s a lesson already: don’t think in terms of your specialty area, think about an aspect of society that could be radically changed.)

—Christina Lomasney of Modumetal will talk about how growing the right kinds of “nanomaterials” could help reinvent the steel industry—and how this will impact energy efficiency, transportation, and defense. She might also tell us a thing or two about what she learned from her time at Boeing.

—Mick Mountz of Kiva Systems (from the Boston area) will discuss how mobile robots are transforming the logistics and operations of warehouses, especially in the e-retail business. Kiva’s robots move shelves of inventory around automatically and are able to help staff distribution centers for companies like Staples and Zappos (now part of Amazon).

—Steve Seitz of the UW’s Department of Computer Science and Engineering will talk about a breakthrough idea from computer vision (and maybe about creating 3-D virtual worlds … Next Page »

Comments (2) | Reprints | Share:          

Whether you’re an entrepreneur, an investor, a researcher, a corporate executive, a journalist, or a patent attorney, you’ve heard it: an idea that will supposedly change the world. In fact, you probably come across someone trying to pitch you an idea like that every week, if not every day. Maybe you even have one you’re working on yourself.

Breakthrough ideas are catalysts of innovation and progress. Think Amazon and its online books, Google and its Web search algorithms, Apple and its iTunes and iPhone. But in the technology and business world, seemingly great ideas are a dime a dozen. For every Amazon, there are dozens of online retailers that have failed. For every Google, dozens of search companies that went nowhere. And so on.

What really counts, of course, is forethought, execution, and results. But that’s much easier said than done. So what makes for a true breakthrough idea? And how can startups and business leaders make their biggest ideas more marketable and scalable, and really change the world? (Thanks to Nick Hanauer of Second Avenue Partners for inspiring this thread in a talk he gave at an NWEN event in December 2008.)

We’re asking these big questions, and we’re going to try to answer them. On the afternoon of March 29, Xconomy is convening a special forum in Seattle called “What’s Your Breakthrough Idea?” The event (agenda and registration info here) will be at the University of Washington, in the atrium of the Paul G. Allen Center for Computer Science & Engineering. We are doing everything we can to make it a must-see for technology, life sciences, and business leaders who have a deep interest in the world’s biggest ideas—and how they will (or won’t) impact the future.

The centerpiece of the afternoon will be an in-depth discussion between two of the area’s top thinkers in innovation across broad realms of science, technology, and society: Nathan Myhrvold, the co-founder and CEO of Bellevue, WA-based Intellectual Ventures, and Leroy Hood, the co-founder and president of Seattle’s Institute for Systems Biology (ISB). They will tackle the important issues of how to think about big ideas, how to tell promising ideas from not-so-promising ones, and how to get maximum bang from ideas that survive. They will speak from experience.

Myhrvold comes from the worlds of physics and software. As Microsoft’s former chief technology officer and the founder of Microsoft Research and later Intellectual Ventures, his company focused on the business of invention, he knows a thing or two about evaluating global-scale ideas. Meanwhile, Hood comes from the worlds of biology, life sciences, and genomics. As the co-founder of ISB as well as a new company focused on early detection of diseases, called Integrated Diagnostics, Hood can speak with great depth on the progress and challenges in biomedicine and human health. Their combined perspectives should make for a tremendously revealing and entertaining cross-disciplinary chat and Q&A.

We’ll also do some deeper dives into specific breakthrough ideas from a select group of speakers, who will include: David Bluhm, CEO of Seattle startup Z2Live, focused on mobile social gaming; Mick Mountz, CEO of the Boston area’s Kiva Systems, a robotics firm working on warehousing and logistics applications; Steve Seitz, professor of computer science and engineering at UW, an expert in computer vision and graphics (he helped develop the technology behind Microsoft’s Photosynth); Dan Weld, UW computer science professor, the co-founder of Netbot, AdRelevance, and Nimble Technology, and a venture partner with Madrona Venture Group; and Norm Wu, CEO of Qliance, a Seattle firm that provides healthcare to thousands while circumventing insurance companies.

Of course, changing the world is never easy. But we’re hoping the discussions and talks on March 29 will inspire our audience to think big, and think realistically—while also making useful connections that otherwise might not happen. Plus we’re going to have an absolute blast with this star-studded lineup, and we think you will too. For more information, and to register for “What’s Your Breakthrough Idea?”, please visit our event page.

Comments (3) | Reprints | Share:          

I had an interesting chat yesterday with Ying Shan, an applied researcher at Microsoft’s adCenter Labs in Redmond. Formed in 2006, the 120-strong lab is dedicated to researching and incubating new digital advertising technologies—primarily to compete with Google’s AdSense in the realm of online contextual ads.

Shan, an expert in computer vision and machine learning, joined Microsoft (NASDAQ: MSFT) last year from Sarnoff in Princeton, NJ. He just got back from chilly Anchorage, AK, where last week he was one of the organizers of a pioneering workshop on “Internet vision”—the melding of computer vision with all the data and technologies of the Web. “This is a really new field,” he says. (The workshop was part of the 2008 IEEE Conference on Computer Vision and Pattern Recognition.)

The Internet vision workshop was notable not only because it was attended by some 800 people, but also because the dozens of participants and committee members represented the major companies that compete (or might want to compete) in the online-ad space: Alex Berg and Tamara Berg from Yahoo Research, Hartmut Neven from Google, Ed Chang from Google China, Eric Hanning Zhou from Amazon, Yingli Tian from IBM Research, Shai Avidan from Adobe Research, Simon Baker from Microsoft Research, Harry Shum from Microsoft… and the list goes on.

A key topic at the meeting (which caught my ear) was how to use computer-vision technology to recognize objects or people in online videos and automatically generate contextual ads. Right now almost all online ads are based on text—extracting keywords from a page, finding the most relevant ads, and putting them up alongside the page. But nobody’s found a good way to do this for video yet. That’s because “image understanding” has been a tough, unsolved technical problem for more than 30 years. But if they could solve this… “That will be a huge, huge thing,” says Shan.

(Some figures: last year, eMarketer projected that online video advertising in the U.S. would grow from $775 million in 2007 to more than $4 billion in 2011.)

So what kind of progress was made at the workshop? Shan gave me a hypothetical example of how video ads might work. Say there’s software that recognizes that Paris Hilton is in a video clip—maybe it does face recognition, matching her to a database that is hand-labeled by users. A company might then put up ads relevant to the celebrity lifestyle, high-end shopping, and so forth. Whoever figures out how to do this reliably, without making many recognition errors, stands to gain a lot.

Some researchers at the workshop showed concrete results in this direction. Xian-Sheng Hua of Microsoft Research Asia presented one piece of research software that used object-recognition algorithms to detect a car in a video clip, and then invoked ads for cars and other vehicles alongside the page. And in an invited talk, Neven from Google showed progress made in recognizing people and places in online photo collections.

Shan stresses that none of these are upcoming products. And judging by the nature of the corporate competitors there at the meeting, I have to believe that anything presented was either very general or at an early-research stage. Nevertheless, it’s a snapshot of the possible future of online ads. “We focused on fundamental issues and basic problems, not too much about specifics,” says Shan. “We try not to ask each other about products and scenarios.”

Comments | Reprints | Share: