AFIS Fingerprint Identification And AI Essay Research — страница 3

an acceptable quality of the print, while taking up only a fraction of the space that a normally encoded JPG file would. This, of course, helps with the speed of the search and keeps the fingerprint database manageable. Fingerprint scan quality can affect the reliability of any electronic fingerprint system. In general, automated fingerprint analysis systems work by creating a computer model of the live print scan. This model is based on many of the features found to be common in fingerprints and is sometimes referred to as a template. The process of creating this model/template is usually referred to as a ‘Registration’ process. Due to a lack of natural moisture in the skin, a dry print can appear broken or incomplete to the electronic imaging system. This can result in

inferior model construction during a registration process or inconsistent matching during a look-up process. The process of matching a live print scan to a model/template is generally referred to as a ‘Lookup’. Dry skin can be caused by a multitude of climatic and environmental conditions. Handling materials or substances tend to absorb or wash the oils from the print. Items such as paper, cloth, wood or chemicals (i.e., acetone, thinners, cleaning agents etc.) will have a direct result on the dryness of the fingers. These items tend to absorb or wash oils from the skin leaving the ridges void of the necessary moisture to reliably electronically image the print. Excessive moisture in the skin can cause line-type features in the print to blend together during the registration

or look-up process resulting in inferior constructs or inconsistent look-ups. An excessively wet print is analogous to viewing a painting after a puddle of paint has been poured on it. Excessive moisture is generally caused by sweating or handling wet materials or substances. Common sources are greasy foods , hand lotion or makeup. The condition is easily solved by removing the excess moisture. Scar tissue has plastic like qualities. When it is dry, it does not image well, when it is wet it looks like a puddle to the imaging system. The AFIS system uses a unique form of ‘Vector Analysis’. It starts by taking a raster scan at an effective resolution of 1000 dots per inch (DPI). Then it makes several passes on the raster data to clean up and optimize the image. This is followed

by a raster to vector conversion process. In short, it intelligently converts raster pixels into vector line types, which are then used to classify the print. Vector analysis generalizes the print through standard classification methods and is therefore very tolerant of micro feature changes or print contamination. This built in tolerance allows the system to be used in applications where other identification systems would be unreliable. The system is trained to analyze and classify a print in the same manner that a fingerprint expert would (only better, easier and faster). The system identifies all Common Line-types (shapes) as described above as well as other pertinent distinguishing marks or relative characteristics of the scanned print/image. In addition, the pattern area can

be classified according to industry accepted rules. All of the analysis is performed by the system and an encrypted print model is generated for optional transmission to a central site for verification. This model (once de-encrypted) represents an access key to the central database for one to one (1:1) or one to many (1:N) lookups. The system consistently maps a scanned print into a fixed coordinate system such that the print always has the same general origin. With this feature, the systems can be used to generate index keys for large existing fingerprint databases. These indexes in turn allow for very fast real time personnel identification.