It’s a little shocking to think about the impact that Microsoft’s Kinect camera has had on the gaming industry at large, let alone the 3D modeling industry. Here at SIGGRAPH 2011, we attended a KinectFusion research talk hosted by Microsoft, where a fascinating new look at real-time 3D reconstruction was detailed. To better appreciate what’s happening here, we’d actually encourage you to hop back and have a gander at our hands-on with PrimeSense’s raw motion sensing hardware from GDC 2010 — for those who’ve forgotten, that very hardware was finally outed as the guts behind what consumers simply know as “Kinect.” The breakthrough wasn’t in how it allowed gamers to control common software titles sans a joystick — the breakthrough was the price. The Kinect took 3D sensing to the mainstream, and moreover, allowed researchers to pick up a commodity product and go absolutely nuts. Turns out, that’s precisely what a smattering of highly intelligent blokes in the UK have done, and they’ve built a new method for reconstructing 3D scenes (read: real-life) in real-time by using a simple Xbox 360 peripheral.

The actual technobabble ran deep — not shocking given the academic nature of the conference — but the demos shown were nothing short of jaw-dropping. There’s no question that this methodology could be used to spark the next generation of gaming interaction and augmented reality, taking a user’s surroundings and making it a live part of the experience. Moreover, game design could be significantly impacted, with live scenes able to be acted out and stored in real-time rather than having to build something frame by frame within an application. According to the presenter, the tech that’s been created here can “extract surface geometry in real-time,” right down to the millimeter level. Of course, the Kinect’s camera and abilities are relatively limited when it comes to resolution; you won’t be building 1080p scenes with a 0 camera, but as CPUs and GPUs become more powerful, there’s nothing stopping this from scaling with the future. Have a peek at the links below if you’re interested in diving deeper — don’t be shocked if you can’t find the exit, though.

Microsoft’s KinectFusion research project offers real-time 3D reconstruction, wild AR possibilities originally appeared on Engadget on Tue, 09 Aug 2011 14:48:00 EDT. Please see our terms for use of feeds.

Permalink Developer Fusion  |  Microsoft Research [PDF]  | Email this | Comments
Engadget

This is from Siggraph 2008. It is a tech talk on the making of Cloverfield. It shows some of the un-rendered scened from the movie and the modeling and rig of Clover itself. Very cool stuff. Visit my Siggraph 2008 Diary for more at www.GoRivera.com

US $16.00 End Date: Friday Feb-10-2012 20:40:30 PST Buy It Now for only: US $16.00 Buy it now | Add to watch list

width="500" height="375" />

Winners of the student contest nabbed the above trophy, plus a $2,000 prize, bragging rights, and special consideration for a spot at Siggraph's 2010's Emerging Technology demos.

(Credit:

Flickr user psychopsyclist

I can tell you my computer password, but unless you type it in exactly the way I do, you won’t be allowed entry. That’s the idea behind Safelock, one of the just-announced winning entries in the UIST 2009 Student Innovation Competition, a Microsoft-sponsored contest aimed at inspiring keyboard innovation. About a month ago, the company sent out prototypes of pressure-sensitive keyboards to 40 international teams, which had four weeks to cobble together their creations. Here are just a few of the cool ideas they came up with:

First place, most useful: Safelock
Safelock, by Jeff Allen and John Howard of Southern Methodist University, biometrically authenticates a user with just eight characters entered. The key (forgive the pun) is that the user has to enter that relatively short password just right. To create a machine-learning algorithm that discovers the unique way each person types, the team measured four keystroke attributes: flight time (the interval between each keystroke); hold time (the amount of time the key was held); maximum pressure; and a curve fit to the pressure over time as a user pressed each key.

The team conducted extensive tests of their system and say it’s “extremely robust.” Says Howard: “99.4 percent of the time, if you’re not me, you’re not able to log into my account.”

First place, most creative: Hidden Forces
This innovation lets users control multiple cursors by waving magnets above the keyboard but not touching it. A four-person team from Carnegie Mellon University accomplished this by placing one small magnet underneath each of the keyboard keys, with the north side facing up.

Julia Schwarz, Brian Lim, Stephen Oney, and Kevin Huang then used a larger magnet (north side facing down) as a cursor. The larger magnet repelled nearby magnets, pushing them against the pressure-sensitive pads and allowing the computer to know where the magnet was located above the keyboard. The innovators were able to control multiple cursors with this technique, turning the keyboard into a multipoint, in-air interaction device.

…

This is utterly insane. A research team from several universities has put together a system whereby you can draw a terrible stick picture sketch, label your blobby objects what they’re supposed to be, and it will essentially photoshop something together for you that meets your criteria. I kid you not. It’s early, experimental, and questionably useful, but it’s just too cool to not share with you guys.

Photosketch uses a sophisticated image filtering system (as well as humans) to pick objects that are what you’ve written and more or less in the position you want them to be. It then takes just the object and pastes it in over the background — whatever you’ve decided that is. It’ll come up with a few composite images, based on different parts, and then you get to choose the one you like. Are you kidding me?

I don’t want to be premature here, but I’d say tentatively that this does appear to be the greatest thing of all time.

They showed it at SIGGRAPH. If this works the way they say it does (and I doubt it does at this point, really), it’d be a party game all to itself. And can you imagine playing with this thing on a tablet? God damn! The future, people!

[via Metafilter]

Windows 7 to bring more happy says latest ad
South Africa’s data network owned by pigeon with 4GB drive attached
FOLD-E! Clothes-folding robot demoed at SIGGRAPH
We talked to the Pirate Party of the United States: Here’s what it’s all about
Flashlight Speed Holster: Sure, why the hell not?

Who wouldn’t want one? Scatter your undies on the ground (if they’re not there already), give FOLD-E the go-ahead, and scant minutes later, there they all are, folded into packets for you. You’ll need mighty small undergarments for this little robot to do his job, though.

In reality, the bot is more a proof of concept; the robot was given instructions in graphical form rather than straight code or instructions, an alternative communication method being put forth by a group at the Japan Science and Technology Agency. The folding mechanism is really just a good way to demonstrate that.

Instead of giving the bot LOGO-like step-by-step instructions (go 20cm forward, grab cloth, move 10cm forward), the robot is programmed to interpret graphical explanations like this one:

There’s more info in this PDF, which summarizes their work. Now they just have to release a big one so my building manager will buy it for the laundry room.

[via Robots.net]