A Cautionary Note about FINGERPRINT ANALYSIS and Reliance on DIGITAL TECHNOLOGY
Cherry, Michael, Imwinkelried, Edward, Judicature
Today the public is acutely aware of the importance of forensic science. Our morning papers regularly carry stories about the role of DNA evidence in both convicting the guilty and exonerating the wrongfully convicted. At night, the popular CSI television programs dramatize the role that forensic experts play in criminal investigations.
Although DNA evidence now attracts the greatest attention, for decades fingerprint analysis was the gold standard in forensic analysis. In fingerprint analysis, an examiner compares two images or representations of the friction ridge patterns on fingers. In the past, in criminal cases, one of the two images, often referred to as the "rolled" image, was typically produced when a person was arrested. As part of the booking process, the police rolled the arrestee's fingertips in ink and then impressed them on a card. The card was subsequently stored in libraries of such cards maintained by local, state, and national government agencies. The data on the cards was classified on the basis of the type of ridge pattern.
The other image, usually termed the "latent," is typically produced at a crime scene. If the police suspect that a criminal might have left a fingerprint impression on a particular surface, such as a glass tabletop, they can use techniques such as the application of special powders to visualize the image. When they find an image, they photograph it for comparison with the images in the library of fingerprint cards.
Again, in the old days, the police used conventional analog cameras and traditional chemical film to take the photographs. If those administering the library could classify the type of skin pattern displayed on the latent, they searched the library for cards of inked images with similar patterns. An examiner then compared the image of the latent to the fingerprint cards with the most similar patterns. Based on that comparison, the examiner might attribute the latent image and the image on a card to the same person. The fingerprint examiner frequently testified about the comparison at trial.
Thus, in the days of yore, living, breathing fingerprint examiners compared the images. Several aspects of that paradigm inspired confidence. To begin with, a human being made a meticulous comparison of the inked and latent impressions. Moreover, that person was working with the best possible images. Admittedly, no image perfectly captures a person's fingerprint pattern, but some are more complete and therefore more reliable than others. For the most part, today, that paradigm is passe. We will not and should not return to the days of yore; but, as we shall see, we need to be far more aware of the pitfalls lurking in the new paradigm.
Computerized fingerprint analysis
To understand the profound differences between the old paradigm and the new reality, we must focus on two questions: who and what. Who conducts the analysis, and what is being analyzed?
Who conducts the analysis-a human being or a computer? During any given year today, government and business must conduct a huge number of fingerprint comparisons. Unfortunately, there are not enough examiners to conduct or verify even 10 percent of the fingerprint analyses that must be completed annually. Assume hypothetically that you have 1,000 experienced, certified examiners who do nothing but fingerprint analysis. If those examiners conduct 20 comparisons a day for 365 days, they will complete only 7,300,000 analyses per year. That number pales in comparison with the number of analyses that must be conducted.
One government agency alone, Homeland Security's US-Visit, has conducted fingerprint searches for over 40 million individuals since March 2005. US-Visit conducts these searches in order to determine (1) whether the arrivee on American soil is the same person who earlier cleared the overseas departure customs and (2) whether that person is on the Watch List.
However, US-Visit is only part of the story. …