PDA

View Full Version : Re: State of the Art in Markerless Motion Tracking



aaguinaldo70
03-12-2008, 02:11 AM
I had the opportunity to observe the Organic Motion system in action at last
year's SIGGRAPH and spoke to one of their engineers a bit. Although I'm
certainly not an expert on their technology, the system is based on
generating a 3D image by mapping pixels of data from multiple cameras. It is
somewhat similar to how optical marker-based systems work in that
intersections of images from multiple cameras triangulate 3D locations of
the subject. Only with this system, image pixels are used instead of
reflective markers. With thousands of pixels being processed, the system can
model small 3D segments as detailed as surface geometry.

It certainly has its uses in animation motion capture, but I also agree with
previous posts that questions remain in regard to its uses in biomechanical
analysis. The main "problem" that is still apparent in current motion
capture systems and that I see would still be an issue with Organic Motion's
markerless technology is the presence of skin or surface motion, regardless
of how precise the system reconstructs 3D images. I believe any "markerless"
system that would be considered advancement in this field would have to
incorporate accurately capturing bone movement without surface interferences
or the need to filter them (see thermal cameras, dynamic MRI). Additionally,
even if skin artifacts are somehow "smoothed," the markerless system would
have to utilize software that can relate these pixels of information to
joint axes and bone segments the way marker sets are designed to do.
However, the company has said the system's data can be imported into Visual
3D so perhaps this already has been addressed.

In any case, studies comparing data from Organic's technology to traditional
marker-based systems would certainly be needed for its use in the
biomechanics arena.

Arnel Aguinaldo, MA, ATC
Director, Center for Human Performance
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


-----Original Message-----
From: * Biomechanics and Movement Science listserver
[mailto:BIOMCH-L@NIC.SURFNET.NL] On Behalf Of Kieran Coghlan
Sent: Tuesday, March 11, 2008 1:06 PM
To: BIOMCH-L@NIC.SURFNET.NL
Subject: Re: [BIOMCH-L] State of the Art in Markerless Motion Tracking

The website for Organic Motion is very impressive, but as far as I can
tell, they don't offer any sort of validation studies for the accuracy
and precision of their 3D motion data. You can download motion data
files (I have not looked at them yet) from their site, but it does not
appear that there are any traditional marker-based data sets to which
one could compare the data from their marker-less capture.

I also am interested to hear from anyone who's used the system, or
better yet, has done a validation study. My initial impression is
that the system may be better suited for the entertainment industry
than it is for the biomedical/life-sciences industry, when compared to
traditional (albeit still flawed) marker-based motion capture
technologies.

Kieran Coghlan BSME, MS
Ph.D. Candidate
Biomechanics Laboratory
University of Southern California

On Tue, Mar 11, 2008 at 7:49 AM, Stefan Litzenberger
wrote:
> Dear List-Members!
>
> I am a laboratory technician at the Fachhochschule Technikum Wien
> University of Applied Sciences, Vienna, Austria.
> To keep my students up to date with the newest developments I am
> interested in all the technology used for motion analysis. As most of
> you might know, markerless tracking is a big issue nowadays. And I have
> read a handfull of interesting papers recently, all concerning problems
> of and solutions for markerless tracking.
> But during my research I found the following:
>
> www.organicmotion.com
>
> A system that provides - as far as I understood - real time (sic!)
> markerless motion capturing with up to 120 frames/sec and a resolution
> of 2 Mpixel.
>
> So: has anyone any experience using this system? Does anyone know of
> similar systems?
>
> I am looking forward to your replies.
> best regards
> Stefan
>
> mit freundlichem Gruß / best regards
> Stefan Litzenberger
>
> --
> DI (FH) Stefan Litzenberger
> Fachhochschule Technikum Wien
> University of Applied Sciences
> Sportgerätetechnik / Sports Equipment Technology
> Höchstädtplatz 5
> A-1200 Vienna
> Austria
> Tel: ++43 1 3334077 - 377
> Fax: ++43 1 3334077 - 369
> mailto:litzenberger@technikum-wien.at
> http://www.technikum-wien.at/
>
> ---------------------------------------------------------------
> Information about BIOMCH-L: http://www.Biomch-L.org
> Archives: http://listserv.surfnet.nl/archives/Biomch-L.html
> ---------------------------------------------------------------
>