Microsoft's markerless game tracking

[Glenn Fleisig]

Hi everyone,



I see that Microsoft has released a game accessory called Kinect
(www.xbox.com/en-US/kinect). This is a full-body markerless motion
capture system. Looking at it, I wonder if this has relevance for the
field of biomechanics.



Is anyone familiar with the technology driving Kinect? How limited is
the tracking accuracy, resolution, and capture speed? Is this relevant
for only 2D or potentially 3D? Will this product lead to advances in
technology for biomechanics?

- Glenn

Glenn S. Fleisig, Ph.D., Research Director
American Sports Medicine Institute
833 St. Vincent's Drive, Suite 100
Birmingham, AL 35205
(email) glennf@asmi.org
(tel) 205-918-2139
www.asmi.org

"Success is not the key to happiness. Happiness is the key to success.
If you love what you are doing, you will be successful."

- Albert Schweitzer

[Glenn Fleisig]

What we know about Microsoft Kinect

Thank you to all who responded about Kinect. Here are some of the
responses I received, which summarize the information:



-----Original Message-----
From: Oevenzohar [mailtoevenzohar@gmail.com]



Hi Glenn

The tech inside kinect originates at an Israeli company called
primesense (http://www.primesense.com). It is in essence 2D with some
depth info through IR strobe ...



Though interesting for the entertainment industry, I think it's accuracy
is insufficient (yet) for medical / biomechanics applications



Cheers

Oshri even Zohar







From: Joseph Gabra [mailto:jgabra@gmail.com]



This may be of interest to you:
http://gizmodo.com/5681078/how-motion-detection-works-in-xbox-kinect


--
Joseph N Gabra

Cleveland Clinic, Lerner Research Institute
Cleveland State University, Graduate Student
jgabra@gmail.com





-----Original Message-----
From: Aguinaldo, Arnel [mailto:aaguinaldo@rchsd.org]





Hi Glenn-



Kinect is primarily based on 3D depth perception through the use of
range camera technology developed by PrimeSense. Using coded data from

an invisible infrared light source, the depth image of a scene can be
mapped out with changes in that depth calibrated for movement. Since

gaming and most mocap applications really only need these movements
"captured" (as opposed to quantified and analyzed), it is an innovative

way to drive the animation without the need of markers or controllers
(ie, Wii remote, PS move). I guess the question for its biomechanical

implications is really how the depth image can be mapped to global and
segmental (body) coordinate systems to allow for kinematic calculations,

but I can see its potential. I believe Organic Motion
(www.organicmotion.com), which is another markerless mocap system, has
developed an SDK for Visual3D to do this.





Arnel Aguinaldo, MA, ATC

Director, Center for Human Performance

Motion Analysis Laboratory

Rady Childrens Hospital

3020 Childrens Way 5054

San Diego, CA 92123

858.966.5424

www.sdchp.com



Assistant Professor, Biomechanics

Department of Exercise and Nutritional Sciences

San Diego State University











From: Jennifer Mandina [mailto:jennifer.mandina@gmail.com]



Hi Glenn,



I am unsure of Kinect's relevance to BME at this point. I can see
possibilities in the future. I own one and played it this weekend. Based
upon my experience, It seems to mostly track in 2D, but has some
understanding of 3D because if you move outside the preferred zone of
the camera it will tell you to step back.



In terms of tracking, most of the games assume that there may be a lag
associated with transfer of information and have the ability to better
attenuate the game to the rate at which information is received. At all
points you can see a little black and white image of what is on camera
in the corner of the screen; however, for the games, you are depicted as
an avatar. The most interesting part of the Kinect software for me was
that when my friend decided to stop playing at the end of the level and
I took over, the software detected that I was a different person and
assigned me a new avatar.



Hope this provides you with some information.



Kind regards,

Jennifer





-----Original Message-----
From: kevin channon [mailto:kc305@cam.ac.uk]



Hi Glenn,



I'm not sure that it is possible to hack the Kinect system yet, but
people are definitely working on it. The open-source software comapny

Adafruit has started the Open Kinect project and offered a $2000 bounty
for the first examples of open source drivers for the system.

You can read more about it at

http://www.adafruit.com/blog/2010/11/04/the-open-kinect-project-the-ok-p
rize-get-1000-bounty-for-kinect-for-xbox-360-open-source-drivers/



So give it a little time and it should be possible to start looking into
this ) If you have the know-how, you could always join in on

the project and help speed things up )



Kevin





From: Glenn Fleisig
Sent: Friday, November 05, 2010 8:17 AM
To: BIOMCH-L@NIC.SURFNET.NL
Subject: Microsoft's markerless game tracking



Hi everyone,



I see that Microsoft has released a game accessory called Kinect
(www.xbox.com/en-US/kinect). This is a full-body markerless motion
capture system. Looking at it, I wonder if this has relevance for the
field of biomechanics.



Is anyone familiar with the technology driving Kinect? How limited is
the tracking accuracy, resolution, and capture speed? Is this relevant
for only 2D or potentially 3D? Will this product lead to advances in
technology for biomechanics?

- Glenn

Glenn S. Fleisig, Ph.D., Research Director
American Sports Medicine Institute
833 St. Vincent's Drive, Suite 100
Birmingham, AL 35205
(email) glennf@asmi.org
(tel) 205-918-2139
www.asmi.org

"Success is not the key to happiness. Happiness is the key to success.
If you love what you are doing, you will be successful."

- Albert Schweitzer