Friday, March 21, 2008

I'll break your kneecaps...

ResearchBlogging.org
Traditionally, the pelvic bone and skull are used in forensic cases and anthropological studies to determine the gender of given skeletal remains. Often, however, full skeletons are not available, and in many cases very few bones at all may have survived. Never fear, however; your kneecaps can give away your sex.
The patella (kneecap) is affected, in shape/size/etc, by the muscle mass of the legs, so it is a relatively safe assumption that these physical attributes would vary within a population; statistically, the patella also displays a sexual dimorphism (it's different for men and women, as are the skull and the pelvis). Studies had shown a relatively good separation of genders, from 67% to 85% in accuracy of determination. A new study out of UT Knoxville has, using a computer-algorithm based "automated feature extraction technique," achieved better than 93% accuracy in gender determination. To describe the very technical method in a very non-technical way, a base "library" of patella CT scans was built up and used to create a "template" kneecap. This allowed them to extract any differences between a given sample patella and the template. The differences were categorized, sorted and ranked, and gender determination (ie, sex classification) could be made using those differences.
So it turns out that even a single kneecap is enough to determine your gender, albeit with a tremendous amount of computing power. I wonder (very tongue in cheek, of course), could one get a discount by signing up for the knee-replacement/sex-change combo?

Reference:
Mahfouz, M., Badawi, A., Merkl, B., Fatah, E.E., Pritchard, E., Kesler, K., Moore, M., Jantz, R. (2006). 3D Statistical Shape Models of Patella for Sex Classification. DOI: 10.1109/IEMBS.2006.259373

2 comments:

  1. Nice! I'll get you the IEEE paper when it is buffed and polished (~1 week). I just clicked over to see what was going on over here, now I understand the facebook comment. ;>

    ReplyDelete
  2. Incidentally, this is just a lame conference paper. Our first journal article will be published in IEEE Transactions on Microwave Theory and Techniques...

    ReplyDelete

Think carefully before you post. I reserve the right to moderate any comments posted to my blog.