Page 34 - Automated Fingerprint Identification Systems (AFIS)

INTRODUCTION 19

once restricted to persons who were fingerprinted as part of an arrest process
or job applications, identification systems, particularly automated systems, are
now used in some very new applications. For example, the U.S. Department of
Homeland Security has piloted a program to perform a fingerprint-based check
on certain foreign visitors as part of the US-VISIT program. Many states now
require that applicants for social service benefits be fingerprinted to ensure
eligibility requirements. More classes of jobseekers are fingerprinted than ever
before, and not only are they fingerprinted and their backgrounds checked,
but also their records are retained for future comparison.
The intentions of policy makers regarding the development of identification
systems may be noble, but their understanding of the issues is often less than
complete and their timelines perhaps unrealistic. There may be, however, a
great deal of public funding to support the development and implementation
of such systems. Government has to rely on the private sector to develop and
bring these technologies to market, and in this arena, are many competing com-
panies. It is also important to realize that the performance results touted by the
marketing department of a company offering an AFIS may be different from
the performance targets developed by the engineers of that company. Too often
policy makers embrace the hype of marketing staff without confirmation from
an outside source. This can result in unrealistic or misunderstood expectations
of the success of the application. And the leap from marketing to newspaper
headlines can ignore even more caveats and limitations. This book removes
some of those gray areas and even provide specific guidelines for improving the
process.
Identification also involves probability and risk. The consequences of a
missed identification or not making an identification vary based on the level of
need for the identification. For example, if the hand geometry reader at a
Disney theme park fails to recognize a legitimate annual pass holder, the pass
holder can walk to the nearby visitor center for assistance. Requiring all pass
holders to have all ten finger images captured in a database may produce a
more accurate identification system, but the cost would be many times greater.
Managers must determine whether the costs of having not identified a few pass
holders or even misidentifying a few persons fraudulently using passes, there-
fore allowing them entry into the park, justify switching to a more secure and
expensive identification process.
This book was written to provide information about the processes that AFIS
systems replaced; it is a history not of fingerprinting but of AFIS. Before auto-
mated identification systems there were semi-automated identification systems.
Beginning in the early 1970s, the introduction of mainframe computers and
punch cards brought what was then state-of-the-art computing horsepower to