Work coming out of MIT’s Media Lab has taken the imaging powers of a Microsoft Kinect and used them to power software that reveals how we’ll be videoconferencing in the future. It’s all about focus, and useful data.
If you’ve ever taken part in a business-oriented videoconference you’ll know all about the problems of people talking off-camera, the difficulty of trying to work out who, at a packed boardroom table, is talking, the distraction-powers of other non-speakers in the room and even the view through another office’s window. While videoconferences are often useful, because they add in a degree of eye-to-eye communication that can be missed in phone-based conferences, they’re far from ideal. And besides, the equipment is often expensive. Work at MIT’s Media Lab, on the other hand, points to Microsoft’s Kinect as a powerful videoconferencing tool and a way to make video-working more commonplace as telecommuting continues to rise.
The approach taken by MIT’s research team takes the idea of a videconference, which is typically just a window into a remote room, into the 21st century–and beyond: “What can we do if the screen in videconferencing rooms can turn into an interactive display?” they asked, taking the remote working notion into a whole new paradigm.
Instantly you’re probably imagining pop-over graphics in an augmented-reality style that tell you who a particular speaker is, who they work for and how much of the meeting has been dominated by their words–a rich analytic graphic that could definitely help unwieldy conferences with a bit more relevance. And MIT’s folks have tackled exactly that idea, with automatic recognition of who’s talking and floating data clouds. But the work goes far beyond that.
How about using Kinect’s object recognition powers to slightly blur everything in the image apart from the person speaking, as a way of adding emphasis and reducing distraction? Or how about removing the entire office background, reducing the video feed to a neat live image of just the people participating? Check and check. What happens if you wish to momentarily duck out of a conference, but don’t wish to be obvious? MIT’s dreamed up a way for you to replace the video feed of yourself with a static image to let you zip off to the restroom or make a vital call. There’s also the ability to add useful data to things you’re demonstrating with an augmented reality angle–perfect for designers or architects perhaps.
MIT’s innovation uses cheap, ubiquitous technology, requiring merely a pre-calibrated conference room and a fast broadband connection. And above all it hints that videoconferencing may be about to get a lot more useful–and probably more fun, too.