Sent to you by khalil via Google Reader:
There have been a number of exciting developments in the last few days related to XBox Kinect. For those that may not be aware of what it is, it's a brand new interface to their Microsoft XBox games console. Kinect brings games without using a controller. By using projected infra-red light, the device is able to map the environment and via xbox built in software recognise gestures of any kind (see video below).
Only few hours after the worldwide release of Kinect, guys at NUI Group, who posted results first, planed to only release the driver as open source once their $10k donation fund was filled up. In the meantime, Hector Martin performed a quick and hack of his own (three hours into the European launch) and released his results and code into the wild – named 'libfreenect'.
Within a few hours of that Theo Watson ported it, making Kinect run from OSX for the first time. Most of Hector's code was left unchanged but with some hacks to the libusb which were needed (and a few extra libusb commands ).
At the same time as Theo was posting the work, Kyle McDonald was working on his point cloud app to make it work with Kinect by just using the saturation value (close range). manually aligning the texture to the depth.
Working together with the rest of the oF community, Kyle, Theo and Joshua Noble, the work they have done has now become an openFrameworks addon and available for download here.
This was quickly followed with openCV blob detection working with ofxKinect, allowing for detection of elements such as hands in real time.
Memo Akten wrote a little demo to analyse the depth map for gestural 3D interaction. One hand to draw in 3D, two hands to rotate the view. Although just an early prototype, very impressive nevertheless. You can download the source for the below demo (GPL v2) here.
Simultaneously, Kyle continued with further point cloud experiments and in his words "you can never have too much DOF" this was the image posted on the forum (download).
and this work by Jason Laan, creating a colour point cloud (download).
Now, using OpenKinect + OpenCV + FestVox + some basic image processing by yankeyan, Kinect is recognising objects in real time on the PC.
Amazing work by the communities and clearly an example of the power of opensource. Daniel Shiffman is now working on the port for Processing and undoubtedly many more are playing with what's already been created. What is also amazing is that this is just the beginning. It's only been few days and some truly amazing work is being created. When it comes to arts/installations and innovative utilisations of Kinect now that different ports are available, makes me wonder how many people will buy Kinect not for their XBoxes but to experiment on their PCs – many for sure, myself included.
/// UPDATE: Nov 15, 2010
With the help of Hector Martin and various members of the openFrameworks community, Daniel Shiffman has just released Processing Kinect library. You can download it here.
Kinect – OpenSource [News] is a post from: CreativeApplications.Net | Follow us on Twitter - Facebook - Flickr - Vimeo
Related Posts:
- 100 Abandoned Artworks [Processing]
- Faceless Void [Processing]
- Brownian Motion [Processing, Inspiration]
- Voltron [openFrameworks, Processing]
- The Root Of The Root [Events]
- Toxiclibs 2010 Showreel [Processing]
Things you can do from here:
- Subscribe to CreativeApplications.Net using Google Reader
- Get started using Google Reader to easily keep up with all your favorite sites