NCJ Number
219471
Journal
Journal of Forensic Identification Volume: 57 Issue: 4 Dated: July/August 2007 Pages: 539-549
Date Published
July 2007
Length
11 pages
Annotation
This study examined the optimum encoding techniques for latent print examiners when conducting searches using the integrated automated fingerprint identification system (IAFIS).
Abstract
Results indicate that although the IAFIS has some inherent difficulties, forensic examiners can overcome them by employing 10 main techniques, including avoiding marking minutiae on the innermost recurve or spike above the shoulders in a loop or whirl; using caution when marking short ridges, islands, and incipient ridges; avoiding encoding outlying minutiae; when certain of the orientation, searching the latent fingerprint as “oriented;” and not guessing in the areas of uncertainty. As expected, the encoding of the total true minutiae provided the highest accuracy. Other findings indicated that unmatched minutiae was more detrimental to the score than unmarked minutiae, the minutiae count alone was not sufficient to determine search reliability, and the IAFIS occasionally had difficulties with multiple clusters of minutiae or minutia outliers. The research focused on the tenprint fingerprint database and the results of latent fingerprint searches. Examiners analyzed sets of 100, 300, and 750 latent fingerprints with multiple encodings, all of which were then categorized as “good,” “bad,” and “ugly.” The examiners assessed tenprint fingerprints on a minutia-by-minutia basis, looking particularly at missed minutiae and false minutiae marked by the automatic feature extractor (AFE). Examiners also assessed difficulties the AFE had for marking minutia types on specific areas of the print. Once the difficulties were noted, the encoding strategy was altered to minimize the problem. The Federal Bureau of Investigation (FBI) continually conducts research on the IAFIS to make it more accurate and user-friendly. Figures