Gamasutra

A new paper from Microsoft Research in Cambridge details how machine learning routines are being used to improve the speed and accuracy of limb-detection techniques for the company's Kinect sensor.

"Real-Time Human Pose Recognition in Parts from Single Depth Images" [PDF, summary] lays out in detail new AI routines that could be used to improve Kinect performance in future software.

The basic process involves analyzing millions of 3D depth maps that were pre-labeled with identifiable body parts -- such as arms, legs and torso. A server consisting of 1,000 cores analyzed roughly a million of these images each day, compiling the aggregate results into a series of trees that could successfully identify the body parts quickly without the identifying data.

Once the trees are fully built, they're used to probabilistically guess a specific body part for each bit of 3D pixel data taken in by Kinect. Finally, the system uses these pixels to assign positions for the joints that make up the skeleton of a character's 3D model.

The researchers claim the Xbox 360 GPU can perform this pattern recognition process in under 5 ms -- a rate of over 200 frames per second -- which the team says is "at least one order of magnitude faster than existing approaches." What's more, the large number of machine learning inputs means the process reportedly works across a variety of different body types without any calibration poses, as seen in this explanatory video.

Early Kinect demos and software received some criticism from developers and observers for a noticeable lag between real-world motions and on-screen reactions when using Kinect, which some measured at hundreds of milliseconds.

http://www.gamasutra.com/view/news/33756/Microsoft_Research_Details_New_Faster_Method_For_Identifying_Player_Bodies_On_Kinect.php?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+GamasutraNews+%28Gamasutra+News%29

Let's hope we see this software on the 360 soon then.

It sounds great, I hope this can be applied via a firmwire update and previously released titles can automatically use this software

I don't know if it's a typo or not, but it says the recognition is done by the GPU, not the CPU.

GPUs can be very effective with computations, so I doubt it's a typo.

Yeah it  could make Kinect lag pretty much non existant.
This is huge news imho.

yes this is something i'm expecting from kinect at the beginning, current kinect is just too gimmick in performance, much looking forward to this... ^^

Perfect! Go ms

I thought it already did this?

How much memory and gpu cycles does it take though?

If it takes up 5ms per frame of the gpu that would suggest 15% time of the gpu is spent on recognition plus overhead from the schedular and swapping data in/out.

They say it is at least one order of magnitude faster then existing approaches. Which means the existing approach an order of magnitude less from 200fps, so runs at less then 20 fps? really?

Just like the wii. It was said on this site a long time ago that these technologies would come about and most likely in a form of a update. November we should see a update in software.