On the psychophysics of matching latents to ten-print patents

For over a century, the matching of images of fingerprints has been used for forensic identification. Despite the extensive forensic history of claimed matches of latent (crime-scene recovered) prints to tenprint standards (full-rolled prints), there have been no published, peer-reviewed studies directly examining the extent to which people can do the task. We report just such an experiment with naive undergraduates using digital scans of real, crime-scene latent prints, and patent prints taken from official, ten-print cards of suspects. The results demonstrate that people can match such crime-scene latent prints quite well to the prints of their corresponding full-rolled ten-print cards, and that matching accuracy varies as a function of source finger type