Answers are grounded in this paper and cite their sources.
Validation study demonstrates examiner response distributions across standardized fingerprint triplets, highlighting utility for forensic calibration and proficiency testing.
The Indiana Fingerprint Comparison Database is an open-source resource consisting of 513 image triplets of fingerprint images. Each triplet was assembled by a subject matter expert and consists of a latent impression, a mated exemplar livescan, and a non-mated exemplar livescan. The images have known ground truth and reflect many of the attributes of casework. A subset of 90 image triplets was tested in a black box study using 104 latent print examiners, and we provide response frequencies for each image pair using a 5-point response scale. We use an ordered probit model [1] to summarize the response distributions as a way of characterizing the strength of support for the same source proposition for each image pair and create likelihood ratios. The black box performance data and the freely-available images described here will support future research studies, laboratory training, validation, internal proficiency testing and calibration exercises.
No takes yet. Share an insight, caveat, or question.
Kibbey et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: