Scientific lineage
Verity did not invent the science of forensic surface comparison — it operationalizes published, open research. The methods below are the shoulders it stands on; Verity’s contribution is the production layer built on top of them. Citation here denotes intellectual lineage only, never affiliation or endorsement.
Automatic bullet-land matching
Verity's bullet-land pipeline descends from the published automatic-matching method — crosscut signatures, alignment, and a learned similarity score — the research line that produced the open bulletxtrctr toolchain. Verity re-implements the ideas in an independent pipeline and benchmarks against the original on the same public data.
Congruent Matching Cells
The cell-counting idea at Verity's core is Song's Congruent Matching Cells, developed at NIST for cartridge-case breech faces. Verity generalizes it to Congruent Matching Regions so one pipeline spans striated and impressed marks, rather than a bespoke method per mark type.
The open-data ecosystem
Verity is validated against the public benchmark data the research community built and shares — the Hamby 10-barrel bullet sets from the open NIST NBTRD, the Fadul consecutively-manufactured cartridge cases from CSAFE-ISU's open cartridgeCaseScans repository, and screwdriver toolmarks from the open tmaRks dataset — and its native X3P codec is interoperable with the community's x3ptools reader.
Not affiliated, and not claiming to be
Verity is an independent project. It is not affiliated with, endorsed by, or developed in partnership with CSAFE, NIST, or Iowa State University. The research cited on this page is public, peer-reviewed work that anyone may build on — building on it implies nothing about its authors’ or their institutions’ views of Verity.
What Verity adds
The lineage above is the science. Verity’s own contribution is deliberately narrow: the deployment, calibration, and audit layer that carries that science into practice.
An open X3P codec
A native Rust reader/writer for X3P (ISO 25178-72) with Python and R bindings — bit-identical I/O across languages.
One calibrated pipeline
Striated and impressed marks flow through a single method that reports a bounded, calibrated likelihood ratio with region-level attribution — not just a score.
Reproducible validation
Source-disjoint, per-dataset evaluation — AUC and Cllr reported without pooling across firearm makes — re-runnable as references grow.
Auditable reports
A per-comparison artifact: the LR, its verbal weight, the regions that drove it, and an explicit statement of scope.
Real, downloadable artifacts
The validation report and the per-comparison report are generated from the live engine on public benchmark data. The technical white paper is authored — the written account of the method behind them.
Built on the field, not against it
Verity does not replace the science or the examiner — it deploys, calibrates, and makes auditable what the research community has published. The full bibliography traces every claim on this site back to the scholarly record.