Counterfeit Listing Detection Network

Detection Controls

Risk threshold 0.55

Crawl / scan rate 6 / s

Price-anomaly weight 1.0×

Live Detector Stats

Scanned

Flagged

Precision

—

Recall

—

Unscanned listing

Cleared (genuine, passed)

Flagged & quarantined (true counterfeit)

Flagged & quarantined (false positive)

How it works

Every marketplace listing carries three synthetic features a real brand-protection crawler would extract: image/logo similarity to the authentic asset s ∈ [0,1], price deviation from MSRP p ∈ [−1,1] (negative = suspiciously cheap), and seller-trust score t ∈ [0,1] (account age, ratings, verification).

risk = σ( 2.6·(1 − s) + w·2.2·max(0, −p) + 1.6·(1 − t) − 2.4 )
flag listing  ⇔  risk ≥ threshold

σ is the logistic sigmoid, so risk is always a probability in (0,1). The simulated crawler pulls unscanned listings from the marketplace at the chosen scan rate, scores each one, and — if risk ≥ threshold — routes it to the quarantine zone for takedown review.

Risk threshold — raising it trades recall for precision, exactly like moving the operating point on a real ROC curve.
Price-anomaly weight (w) — how much a suspiciously low price counts toward risk, independent of how convincing the counterfeit's imagery is.
Precision / Recall — computed against each listing's hidden ground-truth label, the same way a takedown program is graded against confirmed counterfeit reports.

Real-world relevance: this mirrors OSINT brand-monitoring pipelines — automated crawlers score marketplace and social-media listings, and only ones crossing a risk threshold are escalated for a human-reviewed DNS/marketplace takedown request.

scanning…

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.