Gender classification

Jun 22, 2007 07:22

Teaching computers to distinguish between male and female faces is a ticklish problem. Chellappa et al's 1995 review discusses the facial recognition methods used until the mid 1990's, which had limited success. Moghaddam et al (2000) got better results on low resolution thumbnails with their Support Vector machines (SVM) than human test subjects ( Read more... )

faces, gender, rama chellappa, sex differences, technology, gender classification, facial recognition, computers, support vector machines, features, svms, hyun-chul kim, gender differences, baback moghaddam

Leave a comment

Comments 2

astrogeek01 June 22 2007, 14:45:36 UTC
I often have a hard time telling someone's gender. I'm not surprised a computer program would have trouble.

Reply

differenceblog June 22 2007, 14:49:10 UTC
What bothers me is how deeply ingrained the urge to try to classify is. Normally, the classification is so automatic, I don't even notice it happening. Unfortunately, any time it's not fairly clear, I catch myself looking for cues. I've mostly trained myself into looking for voluntary cues instead of strictly biological ones, but I still find myself noticing more than I'm comfortable with.

Reply


Leave a comment

Up