Video: Facial Recognition Software

Facial recognition- “one of the classical and still unsolved problems that has kept computer vision scientists busy since the early 1970s” (Bronstein, 2008). But scientists are developing this new technology in ways entirely new to the computer world. Now-a-days, facial recognition technology can detect or verify a person using video surveillance or digital images from 2-d or 3-d images (Biometrics and facial,” 2008). This can be used for the identification of passport holders, the prevention of people using a fake ID or obtaining a driver license under a false name (Oregon Department of, 2007) and for the detection of possible criminals or terrorists. For example: at the Super Bowl in 2001 police used facial recognition software to detect potential criminals and terrorists and “found 19 people with pending arrest warrants” (Frisch, 2001). Facial recognition software is also used in casinos to detect card counters, thieves and blacklisted people (Bonsor & Johnson, 2001). It’s obvious that this up and coming technology has many benefits and can be highly useful. But how does it work? Where did it come from? And what are its flaws?

“Facial recognition software is based on the ability to recognize a face and then measure the various features of the face” (Bonsor & Johnson, 2001). This application of 3-D facial recognition is done by computers taking what’s called a faceprint, which is a numerical code using the nodal points of the face. Nodal points are the different features of the face, i.e. the peaks and valleys of the face. Some of these landmarks or features are: ‘the distance between the eyes, width of the nose, depth of the eye sockets, the shape of the cheekbones, and the length of the jaw line (Bonsor & Johnson, 2001). Facial recognition is now mostly used with 3-D technology in steps. The first is detection: an image is required by using 2-D photographs or a live video of the person. Once the face is detected, the system aligns it by identifying the head’s position, size, and pose, then measures the nodal points discussed earlier. The measurements are then translated onto a template with unique codes for the individual. Once the system has made all the appropriate measurements it will convert the image to 2-D to find a possible match. The image is then compared to every potential match and one is verified (Bonsor & Johnson, 2001).

Another type of facial recognition is Surface Texture Analysis. It works in a very similar way: first it takes a skinprint which is simply a picture of a patch of skin that is then broken up into smaller blocks and zoomed in on, so to say, so as to identify pores, lines and actual skin texture. If Surface Texture Analysis were to be combined with 3-D facial recognition the identification of an individual could be increased by 5 percent (Bonsor & Johnson, 2001).

In sum, facial recognition systems view an image of a face and compare it to those in its database “comparing structure, shape and proportions of the face; distance between the eyes, nose, mouth and jaw; upper outlines of the eye sockets; the sides of the mouth; location of the nose and eyes; and the area surrounding the check bones.” (“Facial Recognition”).



Biometrics and facial recognition. (2008). Retrieved from http://www.animetrics.com/technology/frapplications.html

Bonsor, K., & Johnson, R. (2001, September 4). How facial recognition systems work. HowStuffWorks.com, Retrieved from http://electronics.howstuffworks.com/gadgets/high-tech-gadgets/facial-recognition.htm#

Bronstein, M. (2008). Expression-invariant 3d face recognition . http://spie.org/x31614.xml?highlight=x2412&ArticleID=x31614 10.1117/2.1200811.1366 .

Facial Recognition. Global Identity Management. Retrieved October4, 2010 from http://www.findbiometrics.com/facial-recognition/.

Frisch, G. (2001, August 20). Privacy is (virtually) dead. Retrieved from http://www.jrnyquist.com/aug20/privacy.htm

Oregon Department of Motor Vehicles, (2007). Facial recognition at dmv Retrieved from http://www.oregon.gov/ODOT/DMV/news/cards_facialrec.shtml

Facial Recognition: Advantages and Disadvantages

Facial recognition technology is a fairly new way of identify people who could be dangerous or need to be located. It works by picking faces out of a crowd, obtaining the measurements necessary and comparing it to the images already in it's database. For more information click here.

Advantages:
- Can prevent card counters, etc. from entering casinos
-Can identify terrorists, criminals, etc.
-Can find missing children
-Prevents voter fraud
-Targets shoppers

Disadvantages:
-Isn't always accurate
-Hindered by glasses, masks, long hair etc.
-Must ask users to have a neutral face when pictures are being taken
-Considered an invasion of privacy to be watched
-Can easily be abused