Biometric Authentication

We also briefly touched on fingerprint authentication. Fingerprint recognition devices
acquire fingerprint images using solid-state capacitance sensing, avoiding the need for
passwords by using minutiae data.
An individual’s finger acts as one of the plates of a capacitor. The other plate consists
of a silicon chip containing a sensor grid array yielding an image. When a finger is
placed on the chip’s surface, the sensor array creates an 8-bit raster-scanned image of
the ridges and valleys of the fingerprint. An analog-to-digital converter digitizes the
array output. These devices can be integrated with smart card authentication platforms.
Many body parts can be used for identification (and hence provide the basis for
authentication and encryption): fingerprints, eyes, facial features, and so on. These
techniques are known as biometric recognition techniques.

The more accurate (that is, robust) the recognition metric, the more processor and
delay overhead are incurred (but also, the more robust the process, the more value it
should have).
Examples of recognition optimization include the following:
 Ridge minutiae recognition algorithms
 Ridge bifurcation and termination mapping (already used in fingerprint search
engines)
 Ridge width and pore and sweat duct distribution