By Karl Ricanek
NOTE: This is an overview of the entire article, which appeared in the September 2013 issue of the Computer magazine.
Click here to read the entire article.
Biometric systems identify a person based on unique physical or behavioral characteristics such as fingerprints, facial features, and gait. When these characteristics are medically altered, identity researchers are challenged. We’ve all heard about criminals who have altered their fingerprints so as not to be identified by law enforcement. The article includes some descriptions of almost unbelievable alterations that have been reported.
Plastic surgery, while widely used to enhance one’s appearance to correct a disfigurement, also presents challenges for biometric systems that rely on facial characteristics such as the relative position, shape, and size of the eyes, nose, and jaw, and cheekbones.
A new challenge to biometric systems derives from the use of sex hormones, primarily estrogen and testosterone, for gender transformation. Hormone replacement therapy (HRT) impacts fat distribution in the face, changing its overall shape and texture.
Fortunately, new algorithm techniques described in the article are being developed to counter these identity-altering medical procedures. Fingerprint modification, plastic surgery, and HRT cutting-edge medical alterations each rely on extensive databases to deal with this emerging dilemma, so the challenge is for researchers to continue to improve the recognition accuracy of these biometric systems.
The article contains illustrations of changes in facial appearance arising from plastic surgery and HRT, and comparative results of various facial recognition algorithms following plastic surgery.
ABOUT THE AUTHOR
Karl Ricanek Jr. (firstname.lastname@example.org), Identity Sciences column editor for Computer, is a professor of computer science and director of the Face Aging Group at the University of North Carolina Wilmington.