NIST logo

Publication Citation: Characterization of Latent Print ,Lights-OutŠ Modes for Automated Fingerprint Identification Systems (AFIS)

NIST Authors in Bold

Author(s): Vladimir N. Dvornychenko; Stephen B. Meagher; Michael D. Garris;
Title: Characterization of Latent Print ,Lights-OutŠ Modes for Automated Fingerprint Identification Systems (AFIS)
Published: December 04, 2013
Abstract: The term ,lights-outŠ for AFIS ten-print fingerprint operations has been around for many years and is generally understood to mean ,no human intervention is involved.Š But is this appropriate for AFIS latent print operations? Is it truly possible to conduct AFIS latent print searches without human intervention? The short answer is: not entirely, but the amount of human expert assistance can be greatly reduced. To determine where the effort can be reduced requires a careful examination of: a) the various stages of a latent print examination and AFIS processes; b) the traditional role of the human expert in each stage; c) the difficulty of automating the task done by the expert; and d) the gain/loss in performance that might result from eliminating human involvement. The results of such an examination are that various AFIS latent print ,lights-outŠ scenarios emerge in a ,tieredŠ manner, though the tiers presented are not necessarily sequential. This paper examines seven scenarios, defined herein as Tiers 1 through 7, entailing ever-increasing amounts of automation. Some of the salient pros and cons of each tier are briefly discussed. It is suggested that more definitive analysis of the seven tiers requires a good cost/benefit model. Part of the benefit of this discussion is to create a common understanding and standard terminology for these scenarios.
Citation: Journal of Forensic Identification
Keywords: Fingerprints, Latent Fingerprints, Latent Prints, Ten-prints, Automated Search, AFIS, AFEM, EFS, Lights-Out, Cost/Benefit Modeling, CSI, Latent Print Expert, Candidate List, Exemplar, True Mate, Identification, Individualization, Exclusion
Research Areas: Biometrics, Computer Forensics, System Development Lifecycle (SDLC)