Kinect technology thread

[Lucid_Dreamer]

First 15 minutes of The Gunstringer for Kinect. It seems a lot of fun and the controls are easy to understand, plus they work well taking into account your character is a marionette. The blend of a videogame and matinee is so appealing to me.


Some parts of the video are hilarious.

That looks like an interesting casual game. I wonder why the devs didn't decide to go cross-platform with this. Maybe it is the same position/situation the devs from Kung Fu Live were in.

 

[AlphaWolf]

published by microsoft would be why

 

[dragonelite]

They should use this kind of control for a space rts

Use that with this

And im sure people will like it.

 

[dragonelite]

haha awesome.

 

[orangpelupa]

microsoft will bring many improvements to kinect?
http://www.eurogamer.net/articles/2011-06-03-how-ms-will-make-xbl-and-kinect-better

 

[(((interference)))]

So is the depth stream from Kinect to the 360 , 640x480 res after all?
Since it appears they can do finger tracking now.

 

[LightHeaven]

So is the depth stream from Kinect to the 360 , 640x480 res after all?
Since it appears they can do finger tracking now.

The finger tracking they have running in kinect sparkle is to detect hand gestures, not tracking the fingers per se... Theoretically they could use the depth data just to point where the hand is and analyze the gesture through the rgb camera (they kinda do this with the facial tracking) if the depth res by itself is not enough...

Shame Kinect Sparkle is 3 dollars, i would love to try if the finger tracking is reliable enough to be used all the time as a click/touch (in KS you only draw when you do the correct hand gesture, it would be nice something like that in fruit ninja kinect for instance)

 

[kagemaru]

KinectFusion demo video. Easily the coolest use of the sensor yet IMO.

We present KinectFusion, a system that takes live depth data from a moving depth camera and in real-time creates high-quality 3D models. The system allows the user to scan a whole room and its contents within seconds. As the space is explored, new views of the scene and objects are revealed and these are fused into a single 3D model. The system continually tracks the 6DOF pose of the camera and rapidly builds a volumetric representation of arbitrary scenes.​

Our technique for tracking is directly suited to the point-based depth data of Kinect, and requires no feature extraction or feature tracking. Once the 3D pose of the camera is known, each depth measurement from the sensor can be integrated into a volumetric representation. We describe the benefits of this representation over mesh-based approaches. In particular, the representation implicitly encodes predictions of the geometry of surfaces within a scene, which can be extracted readily from the volume. As the camera moves through the scene, new depth data can be added or removed from this volumetric representation, continually refining the 3D model acquired. We describe novel GPU-based implementations for both camera tracking and surface reconstruction. These take two well-understood methods from the computer vision and graphics literature as a starting point, defining new instantiations designed specifically for parallelizable GPGPU hardware. This allows for interactive real-time rates that have not previously been demonstrated.​

We demonstrate the interactive possibilities enabled when high-quality 3D models can be acquired in real-time, including: extending multi-touch interactions to arbitrary surfaces; advanced features for augmented reality; real-time physics simulations of the dynamic model; novel methods for segmentation and tracking of scanned objects​

 

[Shifty Geezer]

Now that is incredibly awesome, and what I was anticipating from the tech! It's the best possible augmented reality tech, and I'm sure will do wonders for computer vision.

 

[LightHeaven]

Wow, that is indeed incredibly awesome.

 

[Tap In]

Kinect voice commands benefiting from Microsoft’s Tellme group and Windows Phone Voice

Speech technology from Microsoft’s Tellme group is quickly becoming a bigger part of the company’s products, starting with the upcoming revamps of Xbox Live and Windows Phone. Users will have more opportunities to use voice commands to interact with and control the on-screen experience — listening to text messages in Windows Phone, for example, and responding to them by voice.

...
And people should find the voice recognition to be considerably better than in the past, said Ilya Bukshteyn, a Microsoft Tellme senior director, when we met up today on the company’s Redmond campus.
Here’s why: The technology has been improved by the diversity of voice searches coming in through applications such as Bing on mobile phones. In addition, Microsoft is using a unified, cloud-based service across its different voice applications. With a larger collection of data to work from, the unified system can learn more quickly.
“We’ve seen more improvement in the last 18 months to two years than we saw in a decade before that,” Bukshteyn said

 

[Scott_Arm]

When does the new dashboard come out?! I've never seen it! The voice integration is very cool.

 

[egoless]

Delete

 

[ERP]

To my mind this is where MS has a potentially huge advantage with new input devices and similar fringe technologies, they have so much existing research and in some cases production technology they can leverage.
How well they actually do that remains to be seen, but a Sony or a Nintendo is not going to invest as much in say image recognition or voice recognition as MS already did before Kinect was even an idea.

 

[Arwin]

To my mind this is where MS has a potentially huge advantage with new input devices and similar fringe technologies, they have so much existing research and in some cases production technology they can leverage.
How well they actually do that remains to be seen, but a Sony or a Nintendo is not going to invest as much in say image recognition or voice recognition as MS already did before Kinect was even an idea.

Surely at least image recognition is a department where Sony is doing a lot already in its camera divisions? My WX10 camera and its companion software (PMB) have a fair bit of image recognition stuff built in.

 

[patsu]

I was told Nuance is king of voice recognition.

MS should be king of depth-based image recognition. Sony's imaging division should be pretty impressive (Aibo was interesting although some other robots did computer vision using 2 cameras [a la 3D vision]).

I bet IBM has a few surprises too... like their Watson project for voice tech.

EDIT: If I remember correctly, Aibo also used ultrasonic tech on top of "vision".

 

[Shifty Geezer]

Surely at least image recognition is a department where Sony is doing a lot already in its camera divisions?

I can't see much relevance for Sony in the 3D space. Sure, 2D image analysis for exposure and tracking makes sense, but they have no need to try and reconstruct 3D worlds. I wonder if anyone's been trying this seeing as 3D cameras are so new? MS's research is much more generalised, without a specific purpose unlike Sony's CE devision, and they'll be looking at what a computer can do long before they determine a real application for it.

The obvious limitation with integrating this latest tech with Kinect on 360 is the requirement for the camera to be moving and exploring the room. The real target for this tech will be 3D cameras on a head-display, overlaying AR components into the real world. But an improved EyePet type game would be truly amazing. Set that up in stores and it'd boggle people's minds!

 

[ERP]

Surely at least image recognition is a department where Sony is doing a lot already in its camera divisions? My WX10 camera and its companion software (PMB) have a fair bit of image recognition stuff built in.

It's the way MS does research and the scope of the investment that sets them apart.
MS is the only company I know of that does research the same way that academic institutions do research i.e. not directed at a product.
And they have a much larger research investment than any of the other players in the console space. Even if it is a bit smaller than it was pre layoff.

Where MS does fail sometimes is getting the research into an actual product, there are a number of reasons why this is the case, but when it does work it's generally because product groups take existing work and commoditize it. The games group seems to be pretty pro active in this area.

 

[Alucardx23]

How Microsoft engineered Kinect to withstand gamers and lightning strikes
http://venturebeat.com/2011/08/19/h...ct-to-withstand-gamers-and-lightning-strikes/

 

[phenix]

Kinect 2 so accurate it can lip read

http://www.eurogamer.net/articles/2011-11-25-kinect-2-so-accurate-it-can-lip-read