FlashID
Match the handwriting. Surface the author.
FlashID is a handwriting analysis, matching, and visualization tool that helps law enforcement, military, and intelligence examiners compare questioned documents against known writer samples — across documents, languages, and scripts.
Threat actors don't always type.
Latin-only tools stop at the language line.
Manuscripts, ledgers, and intercepted notes pile up in evidence rooms because the standard handwriting tooling can't span the document trail your operators actually encounter.
-
Manual side-by-side review doesn't scaleExpert pair-wise comparison takes hours per document. As caseloads grow, backlogs grow with them — faster than examiner hours can keep up.
-
Markup, tracing, and per-character setupTraditional systems require examiners to mark up or trace samples before analysis even starts — pure overhead before any real comparison happens.
-
Subjective scoring that doesn't survive courtOpinion-based conclusions don't translate cleanly across examiners.
-
Latin-only tooling silently failsMost discriminators assume Latin script. The Arabic, Chinese, and mixed-script documents your team actually encounters get rejected, or worse, scored wrong.
One workflow, every script you care about.
FlashID does not require manual markup or tracing. Scan and compare — it works with handwritten and hand-printed text in Latin, Arabic, and Chinese character sets, and surfaces the writer with industry-leading discriminator and comparator algorithms.
-
Scan and compare — no markup, no tracingRoughly 200 per-sample measurements, automatically selected by Linear Discriminant Analysis. Operators don't think about it after scanning.
-
Works on as little as ¼ page of writingQuestioned documents match against references of just one page — real-world evidence sizes, not laboratory-perfect samples.
-
Peer-reviewed, court-defensibleValidated in Journal of Forensic Sciences and Forensic Science International, with an independent CSAFE study and an FBI Laboratory Division feasibility deliverable.
-
Tested on 20+ scriptsHandwritten and hand-printed text in Latin, Arabic, Chinese, and more — already validated across more than twenty languages.
Innovative techniques and approachs to aid investigations.
Competitive Matrix scoring
FlashID tests the questioned document against pair-wise comparisons between every writer in the database and a fixed “test bed” of writers. Works on questioned documents as small as ¼ page and reference samples as small as one page — no markup or character tracing required.
Side-by-side & heat map
A comprehensive suite of visualization tools helps the examiner make their determination — including side-by-side document viewing and an innovative “heat map” of writing similarity.
Grapheme-level graph matching
Reference documents are segmented into words, characters, and basic graphical forms called “graphemes.” Graph-matching algorithms compare them first by topology — edges, vertices, and connectivity — then by geometric features like the shape of curves, angles, and distances.
Built around how examiners actually work.
Easy to use
No manual markup or tracing. Scan the questioned document, scan the known sample, and compare.
Sophisticated algorithms
An industry-leading handwriting discriminator paired with an advanced visual comparator that scores questioned handwriting against known samples.
Multilingual support
Works with handwritten and hand-printed text in any language using Latin, Arabic, and Chinese character sets.
Accelerating the time from unknown writer to potential matches
Peer-reviewed research underpinning FlashID.
Published in the Journal of Forensic Sciences, Forensic Science International, and delivered to the FBI Laboratory Division — the science behind the matcher.
Elucidating the relationships between two automated handwriting feature quantification systems
Fuglsby, Saunders, Ommen, Buscaglia, Caligiuri. Over 81,000 pairwise scores compared between FLASH ID and the kinematic system MovAlyzeR — the static, image-based features in FLASH ID track closely with kinematic pen-stroke features.
Read paper →Comparing handwriter and FLASH ID, Two Handwriting Analysis Programs
Reinders, Ommen, Carriquiry (Center for Statistics and Applications in Forensic Evidence). An independent side-by-side evaluation of the open-source handwriter package and FLASH ID on common handwriting tasks.
FLASH ID — Objective Scoring Feasibility
Sciometrics deliverable to the FBI Laboratory Division evaluating the feasibility of objective, automated scoring in FLASH ID for forensic handwriting identification.
Read report →A Set of Handwriting Features for Use in Automated Writer Identification
Miller, Patterson, Gantz, Saunders, Walch, Buscaglia. Describes the graph-based system at the core of FLASH ID — segmenting handwriting into graphemes, skeletonizing them, and matching first by topology and then by geometric features.
Read paper →Score-based likelihood ratios for handwriting evidence
Hepler, Saunders, Davis, Buscaglia. Compares three score-based likelihood-ratio constructions on cursive samples from over 400 writers — showing how subtle choices in denominator-database generation drive meaningfully different forensic inferences.
Read paper →Using Automated Comparisons to Quantify Handwriting Individuality
Saunders, Davis, Buscaglia. Uses an automated comparison procedure on a database of handwriting samples to empirically bound the random-match probability between two writers with indistinguishable writing profiles.
Read paper →What teams ask before they pilot.
How does FlashID work?
How can FlashID help me?
-
Triage massive data volumes:
Instantly sort and filter large batches of handwritten documents to isolate high-interest files
without manual markup.
-
Conduct rapid 1-v-1 comparisons:
Run quick, automated checks between two documents to immediately score their similarity and decide if a deep-dive forensic review is necessary.
-
Track high-value targets:
Search unknown handwriting samples across your entire database to uncover hidden aliases or identify similar writings belonging to a specific high-value individual.
-
Cross-link biometric identities:
Connect writing habits to other biometric markers—such as identifying a specific threat actor and linking their handwriting profile to established fingerprint or facial data.
Will FlashID replace Forensic Document Examiners?
What is the recommended hardware to run FlashID?
For a Desktop:
- Dual Processors (12 core/24 thread) or better (Example: Dual Intel Xeon)
- 128 GB of RAM
- 2 TB SSD or better
- Windows 11 Pro
- Processor (16 core/16 thread) or better (Example: Intel Core Ultra 9 285H)
- 64 GB of RAM or better
- 2 TB SSD or better
- Windows 11 Pro
See FlashID work on the documents you bring.
Demos, free trials, and pricing — usually within a week of a first call.