Smart Computing Article - fab to FASTERnet.org

One of the four basic face recognition technologies is Eigenface, which means "one's own face." It employs 2-D grayscale images that represent distinctive features. Eigenface technology is patented by MIT (Massachusetts Institute of Technology). Image courtesy of Baback Moghaddam. Face recognition is the biometric technique of identifying individuals by facial features. Face recognition is a type of biometric analysis, which aims to identify people by using biological or anatomical human characteristics, such as fingerprints and handprints, iris patterns, voice patterns, and facial structure. More specifically, face recognition employs digital video cameras and software to capture an image, isolate a human face, and extract specific features of the face that do not change. Such features may be upper sections of the eye sockets, the area around the cheekbones, and the sides of the mouth. The individual's features are then compared to other feature-sets in a database of photographs. Face-recognition technology works with off-the-shelf PCs and video cameras, making this identification technology relatively inexpensive to install and operate. High-resolution images provide even better matches. There are four basic methods employed in face-recognition software: eigenface, feature analysis, neural network, and automatic face processing. Eigenface, which means "one's own face," employs 2-D grayscale images that represent distinctive features. MIT (Massachusetts Institute of Technology) patented this technology. Feature analysis is similar to Eigenface, but can better accommodate changes in facial movements, such as smiles and frowns. Neural network technology uses both the newly scanned face and the stored face to determine if there is a match. An algorithm looks at the similarity of unique features. Lastly, automatic face processing uses distances and distance ratios between obvious features such as eyes, the end of the nose, and corners of the mouth. Although it is an ideal application for security, the use of face recognition for public security