In the last decade especially after 9/11 terrorist attacks, due to rising security concerns, “biometrics”
which was once a highly theoretical research topic, has become a part of our daily lives. Biometrics is a general
term either to describe a process or a characteristic.
As a characteristic, it is a biological (anatomical or physiological) or behavioral characteristic that can
be used for automated recognition.
Iris, fingerprint, hand, face, voice, retina, DNA, odor, lips are some of
biological biometric characteristics whereas signature, keystroke, gait and voice can be listed as behavioral
As a process biometrics defines automated methods of recognizing and classifying individuals based on
measurable biological or behavioral characteristics.
A typical biometric system has five constituents: sensor, signal processing algorithm, data storage,
matching algorithm and decision process. Firstly data is collected by a sensor and converted to digital format.
Then, a signal processing algorithm performs quality control and develops biometric template. This biometric
template is compared with other templates kept in data storage via matching algorithm. Finally, a decision
process uses results from matching algorithm to make a final decision.
Commonly implemented biometric modalities include fingerprint, face, voice, iris and hand geometry.
Early facial recognition techniques used simple geometric models, but latest developments in facial
recognition algorithm put this technology into spotlight.
A general statement of the facial recognition problem
(in computer vision) can be formulated as follows: Given still or video images of a scene, identify or verify one
or more persons in the scene using a stored database of faces.
In facial recognition systems, facial image which is acquired from various forms of media is separated
with complicated background and then normalized through the position of eyes, nose and mouth.