Feature

Document Fraud & Spoof Detection

Catch fake and tampered IDs the moment they're captured. Multi-layer image forensics flag screen-replay, printed copies and recompressed forgeries during the video call — with the exact reasons surfaced to your agent.

Overview

A Real Document, Not a Photo of One

The easiest way to defeat KYC isn't a master forgery — it's holding up a phone showing someone else's Aadhaar, or a freshly printed colour copy. BASEKYC inspects every captured document with layered image forensics that look for the tell-tale fingerprints of a fake: the moiré interference of a photographed screen, the recompression ghosts left by edited or re-saved images, missing or inconsistent camera metadata, and the unnatural edges and glare of a printout. Each document gets a verdict, a confidence score and a plain-language list of reasons, surfaced to the agent the instant the image is captured and stored against the case for later review. Roll it out in warn-only mode to tune for your customer base, then switch to enforce when you're ready to block suspicious uploads outright.

Key Highlights

  • Detects screen-replay (phone/monitor) and printed-copy document spoofs
  • Multi-layer forensics: EXIF, moiré, JPEG-ghost and edge / specularity analysis
  • Real-time verdict with confidence score and human-readable reasons
  • Warn-only or enforce modes — collect data, then block when ready
  • Verdicts stored per document for audit, review and threshold tuning

Capabilities

What You Can Do

Screen-Replay Detection

Moiré-pattern analysis (FFT) catches the periodic interference left when a customer photographs an ID displayed on a phone or monitor instead of presenting the physical document — the most common video-KYC attack.

Print & Recompression Checks

Edge and specularity analysis flags the flat texture and glare of a printed copy, while JPEG-ghost detection exposes regions that were edited or re-saved — the signature of a digitally tampered or recompressed forgery.

Metadata & EXIF Forensics

Missing, stripped or inconsistent camera metadata is a strong signal an image didn't come straight from a live camera. EXIF inspection adds a cheap, high-signal layer on top of the pixel-level checks.

Explainable Verdicts

Every document returns a real / not-real verdict, a confidence score and per-layer scores with a list of reasons in plain language — so agents and auditors understand exactly why an image was flagged, not just that it was.

Warn-Only or Enforce

Start in warn-only mode: verdicts are logged and shown but never block a capture, letting you measure false positives against real traffic. Flip to enforce to automatically reject suspicious uploads once thresholds are dialled in.

Layered With Liveness & Face Match

Document spoof detection sits alongside live liveness challenges and face-match scoring, so a fraudster has to beat presentation-attack checks on both the person and the paperwork in the same session.

How It Works

01

Document Captured

During the video call the agent captures the customer's PAN, Aadhaar and photo. Each image is sent for analysis without slowing the capture — the check runs in parallel and never blocks a successful upload.

02

Forensics Run

The image passes through the EXIF, moiré, JPEG-ghost and edge/specularity layers. Each returns a score; together they produce a verdict, an overall confidence and a list of reasons explaining the call.

03

Flag, Store & Decide

A suspicious result raises a badge for the agent and is stored against the document. In warn-only mode the session continues for review; in enforce mode the upload is rejected on the spot.

Compliance

Fraud Controls Regulators Expect

RBI V-CIP Anti-Spoofing

The RBI Master Direction on KYC requires that V-CIP processes incorporate liveness checks and safeguards against impersonation and the use of forged or photographed documents. Document spoof detection directly addresses the presentation-attack vector regulators call out for live, agent-led verification.

PMLA Customer Due Diligence

Under the PMLA framework, regulated entities must take reasonable steps to verify the genuineness of identity documents during customer due diligence. Automated forensic checks raise the bar against fabricated KYC and create a defensible record that diligence was performed.

Evidence for Every Decision

Each verdict — score, layer breakdown and reasons — is stored against the document and folds into the case audit trail. If a verification is ever challenged, you can show precisely what was checked and why a document was accepted or rejected.

Risk-Based Configuration

Warn-only and enforce modes let you align controls with RBI's risk-based approach — tighter enforcement for high-risk products and customer segments, looser thresholds where friction must be minimised, all without code changes.

Specifications

Technical Specifications

4-Layer

Forensic Pipeline

Moiré FFT

Screen-Replay Signal

JPEG-Ghost

Tamper / Recompress

EXIF

Metadata Inspection

Confidence

Score + Reasons

Warn / Enforce

Configurable Mode

Non-Blocking

Fire-and-Forget Check

Per-Doc

Stored Verdict

Related Use Cases

Further Reading

Stop Fakes at the Source

See how layered document forensics catch screen-replay and printed-copy spoofs in real time, with explainable reasons for every flag.