Verity
Open-science forensic surface comparison

Verity

One calibrated, explainable method for comparing forensic surface marks — a likelihood ratio with a characterized cost, and a map of exactly which regions drove the match.

Try a comparison →The science
Striated · bullets & toolmarksImpressed · cartridge casesCalibrated LR + CllrRegion attribution

The scientific contribution

Two decades of review — the National Research Council’s 2009 report and the 2016 PCAST report — found that most pattern-evidence disciplines lack demonstrated foundational validity, and Cuellar et al. (2024) showed that no discipline has a characterized error rate. Existing tools are bespoke to one mark type and stop at an uncalibrated score. Verity answers this directly: a single, transparent method that reports a calibrated likelihood ratio with a quantified cost (Cllr) and region-level attribution, spanning striated marks (bullets, toolmarks) and impressed marks (cartridge breech faces) by generalizing the Congruent Matching Cells method (Song, 2013) to arbitrary toolmarks. The representation only produces a score; the reportable decision is a monotone, bounded likelihood ratio — auditable no matter how the score was computed.

  • National Research Council (2009) — Strengthening Forensic Science in the United States: A Path Forward.
  • PCAST (2016) — Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods.
  • Cuellar, Vanderplas, Luby & Rosenblum (2024) — Methodological problems in every black-box study of forensic firearm comparisons.
  • Song (2013) — Proposed congruent matching cells (CMC) method for ballistic identification.
  • Brümmer & du Preez (2006) — Application-independent evaluation of speaker detection — the Cllr cost.

Compare two marks

Each mark is a bullet — select all of its land scans (e.g. 6). A single land is only weakly diagnostic; the strength comes from aggregating the lands.