Cross-signal counterfeit detection

See counterfeiting before it spreads.

Individual scan anomalies are noise. Patterns are signal. The counterfeit radar aggregates failed photo-authentications and scan anomalies geographically and over time into scored threat clusters — surfacing cloned QRs and regional counterfeiting that single events never reveal.

Cross-signal clusteringGeographic + temporal windowsScored severity + alertsRuns daily, no PostGIS

2 signals

Combined

Photo-match + scan

~55 km

Geo cells

Grid bucketing

Daily

Recompute

Rolling window

3 tiers

Severity

Low / medium / high

Why clusters, not single events

One odd scan is noise. Twenty in a region is a factory.

A single impossible-travel scan or one failed photo match can be innocent. The radar finds the concentration — the same product failing visual checks across a region, or a certificate scanned from many places at once — and scores it.

SIGNALS

Two independent fraud signals.

The radar fuses failed visual authentications (consumer photos that don't match the certified object) with scan anomalies already detected per-scan: impossible travel and velocity spikes. Each contributes weighted suspicion.

Failed photo-matches

Every geolocated no-match / weak-match from visual authentication is a strong counterfeit signal — the object itself didn't match.

Impossible travel

The same certificate scanned in distant places faster than physically possible flags a cloned QR.

Velocity spikes

A burst of scans on one certificate in minutes signals bots or a duplicated code.

CLUSTER

Aggregated by place and time, then scored.

Events are bucketed into geographic cells over a rolling window per certificate. A cell that accumulates enough weighted suspicion becomes a cluster with a confidence score and severity — no PostGIS required, reusing the platform's haversine math.

Geo-cell + window aggregation

Suspicious events are grouped into ~0.5° cells over a multi-day window. Enough concentration promotes a cell to a cluster.

Weighted score + severity

No-match weighs heavier than weak; high-severity anomalies weigh more. The total maps to low / medium / high.

ACT

Alerts, triage, and the full radar.

When a cluster first appears at medium or high — or escalates to high — the producer is alerted in-app. The Counterfeit Radar dashboard ranks live clusters with their signal breakdown for triage.

Proactive alerts

A new medium/high cluster raises an in-app notification with location and score, so you act before it spreads.

Triage workflow

Dismiss false positives or mark clusters resolved; the radar tracks open threats by severity.

One radar, every signal

Clusters sit alongside B2B partner rejections on the Counterfeit Radar, for a single counterfeit view.

Use cases

Real scenarios where this service delivers.

Cloned QR detection

A single certificate scanned from many distant locations in a short window surfaces as a high-severity cluster — a copied code in the wild.

Regional counterfeiting

A product repeatedly failing photo authentication in one region reveals a local fake operation.

Channel monitoring

Clusters next to B2B partner rejections show where in the channel fraud concentrates.

Why Fidaty

Built for brands that take trust seriously.

Every service in the Fidaty platform is integrated with the rest: identity, compliance, analytics and consumer engagement flow seamlessly. No data silos, no duplicate work.

Book a demo
  • Combines visual-match failures with scan anomalies in one score
  • Finds geographic clusters single events can't reveal
  • Proactive in-app alerts on new medium/high clusters
  • Triage workflow with severity tracking
  • Runs daily as an aggregate job — no PostGIS, low overhead
  • Unified with B2B rejection radar in one dashboard
Verified on-chain

Turn scan noise into counterfeit intelligence.

Let Fidaty cluster your fraud signals geographically and alert you the moment a counterfeit pattern emerges.