Fake Passport Detection: 8 Forensic Signals That Expose Travel Document Fraud

The machine-readable zone at the bottom of a passport's biodata page encodes name, nationality, date of birth, and document number in a strictly formatted string with built-in check digits. Automated verification recalculates these check digits and cross-references MRZ data against the visual fields above.

Fraudsters who alter a field visually but forget to update the MRZ — or who generate an MRZ with incorrect check digits — are caught immediately. This is one of the most reliable signals available.

Photo substitution attacks replace the portrait zone of a genuine document. The boundary between the original document substrate and the inserted photograph shows detectable differences in:

• Sharpness gradients at the portrait edge
• Compression ratio differences between portrait and surrounding content
• Lighting direction inconsistency (the face lit from a different angle than the document background)

Genuine passports include holographic overlays with characteristic optical properties: iridescence, angular colour shift, and fine microstructure. Digital representations of these features — even in high-resolution scans — show different spectral properties than genuine holograms. Analysis of hue variance across the hologram region detects flat digital overlays that simulate holographic appearance.

Every passport printing system produces character shapes with consistent metrics — letter spacing, stroke width, and baseline alignment. When individual characters or fields are replaced (altering a birth year, for example), the replacement characters typically differ at the sub-pixel level from surrounding text, even when the same font is used.

AI-generated faces and document backgrounds have characteristic frequency-domain signatures. When a diffusion model or GAN generates an image, the spatial frequency distribution differs from photographs of physical objects. Spectral analysis of the document image detects these AI generation fingerprints, flagging synthetically produced identity documents regardless of how visually convincing they appear.

The fine-line security background (guilloche) in passports is extremely difficult to reproduce faithfully at print resolution. Digital forgeries either omit the guilloche, use a low-fidelity approximation, or reproduce it with visible rasterisation artefacts. Frequency-domain analysis of the background region detects these deviations from genuine security printing.

Semantic analysis cross-references stated dates (birth, expiry, issue) against document version and issuing country format. A passport with an issue date inconsistent with the version series, or an expiry gap that doesn't match the country's standard validity period, is flagged for review regardless of whether any visual tampering is detected.

For documents submitted as PDFs or digital files, metadata analysis checks creation software, modification history, and embedded color profiles. A "scanned passport" that was actually generated as a native PDF, or one whose metadata shows editing after the scan date, carries strong provenance signals of fabrication.