How We Can Improve the Reliability of Fingerprint Identification
Cherry, Michael, Imwinkelried, Edward, Judicature
Fingerprints are an excellent source of unique identification. Unfortunately, enormous population growth, worldwide terrorism, and the incorporation of new fingerprint features have occurred without the benefit of badly needed reengineering and recalibration. Our fingerprint matches are no longer foolproof.
A perfect storm is developing. The reliability of fingerprint identification has declined while the population of fingerprints has exploded. Who is going to tell prospective employees that their fingerprints indicate they have a criminal background? Every night an automated process takes place where the fingerprints of approximately 50,000 job applicants are computer analyzed as a part of civil and government pre-employment investigations. Unfortunately, the FBI systems performing the fingerprint screenings have significant calibration issues that can affect decisions of great importance, both civil and criminal, domestic and international.
Our fingerprint concerns used to be local. Local police departments and agencies maintained their own collections of fingerprints and there was little coordination among those with fingerprint databases. Now, with the threat of international terrorism, the United States faces a global population of dangerous individuals. National security requires a national system that will enable us to correctly identify fingerprints. If analyzed properly, fingerprints can be as accurate as DNA.
The intent of this article is to make several fingerprint identification recommendations to correct our fingerprint problems and reestablish reliability. It describes some old and new concepts regarding fingerprints, briefly traces the evolution of fingerprint examination systems, identifies the major problems in the current system, and offers recommendations for addressing those problems.
Nature appears to be random until you examine it closely. There are approximately six billion individuals in the world. However, most of them have dark hair, dark eyes, are right handed, and share one of two major blood types. If we are to reliably identify criminals and terrorists, we need to rely on their other physical characteristics. Of course, one such characteristic is DNA type. In truth, though, fingerprints offer amazing uniqueness. A full fingerprint is unique and even if the probability of two independent people accidentally having the same fingerprints is only 1,000,000 to 1, then under the same assumptions the probability of the same two people having not one but three matched fingerprints would be 100,000,000,000,000,000 to 1.
In the late 19th century Sir Francis Gallon developed the first system for classifying and identifying fingerprints. He is quoted as having said that "the odds of two individual fingerprints being the same are 1 in 64 billion." The world population exceeds six billion, and each person has 10 fingerprints. Thus, the world population of fingerprints is now close to 64 billion. If we accept Galton's assertion, there may well be instances of different persons possessing the same single fingerprint. To ensure reliability, we then need to take into account an individual's neighboring fingerprints.
Fingerprint identification is predominately based on viewing all the ridges within a fingerprint and then categorizing it into one of three somewhat similar appearing patterns-loops, arches, and whorls. Each of these three patterns can be sub-divided into one of several sub-patterns. The final step is to find and map the location of small predetermined shapes and contours.
When fingerprints are said to "match," it means the pattern, subpattern, and at least some of the small predetermined shapes and contours present roughly correspond with each other. Since it is very difficult for humans to visualize all of the small predetermined shapes and contours that are present in a fingerprint, a rigorous, systematic procedure is required. …