Visual Inspection Zone (VIZ)
The visual inspection zone, usually shortened to VIZ, is everything on an identity document’s data page except the machine-readable zone. Portrait, field labels, printed values, issuing authority, security features — all of it. ICAO Doc 9303 defines it as the area intended for human reading, which is precisely what makes it difficult for machines.
| Abbreviation | VIZ |
| Defined in | ICAO Doc 9303 |
| Contains | Portrait, name, date of birth, nationality, dates, issuing authority, signature, security features |
| Intended reader | Humans |
| Layout | Varies substantially by issuing state |
| Field labels | Often in the national language plus English or French |
| Counterpart | The machine-readable zone |
Why the VIZ is the hard half
The MRZ is trivially parseable because the standard fixes everything: character positions, typeface, field order, length. Count characters and you have the data.
The VIZ fixes almost nothing. Doc 9303 specifies which data elements must appear, not where they sit, what typeface renders them, what language labels them, or what background artwork sits behind them. A German ID card, a Brazilian passport, and a Japanese passport all satisfy the same requirement with completely different layouts.
So reading the VIZ is a genuinely harder computer vision problem than reading the MRZ. There is no positional shortcut. The system has to recognize the document type first, then apply a layout model specific to that issuer and that design generation, then read text rendered in an arbitrary typeface over decorative security printing.
Why the VIZ matters for identity verification
It would be reasonable to ask why anyone bothers, given the MRZ is right there and easy.
Three reasons. The portrait lives in the VIZ, and it is the biometric reference the whole identity check depends on. Several fields appear only in the VIZ — place of birth, address, issuing authority, and various national additions are frequently absent from the MRZ entirely. And most importantly, reading both zones creates the cross-check.
The MRZ and the VIZ encode the same core facts independently. When someone alters a genuine document, they typically change the printed values, because that is what a human border officer looks at. The MRZ often goes untouched, or gets recomputed incorrectly. That disagreement is one of the highest-value fraud signals available, and capturing it requires reading the difficult zone as well as the easy one — which is why identity document verification cannot stop at the MRZ.
VIZ vs MRZ
| Visual inspection zone | Machine-readable zone | |
|---|---|---|
| Intended reader | Humans | Machines |
| Layout | Varies by issuer and design generation | Fixed character positions |
| Typeface | Issuer’s choice | OCR-B, Size I |
| Error detection | None | Check digits |
| Contains the portrait | Yes | No |
| Reading difficulty | High — needs layout models per document | Low — positional parsing |
What the VIZ can’t do
It has no error detection. Nothing in the VIZ corresponds to check digits. A misread name or transposed date has no arithmetic to catch it, so confidence has to come from image quality and cross-checking instead.
It is the easiest zone to alter. Printed values on a laminated page can be modified with modest equipment. This is exactly why VIZ-against-MRZ comparison matters.
Layout models age. Every new design generation an issuer releases needs new handling, and old designs stay in circulation for a decade. Coverage is continuous maintenance, not a one-time build.
Reading it proves nothing on its own. Extracting the text says what the document claims. Whether the document is genuine depends on its security features and, where present, its chip.
Frequently asked questions
What does VIZ stand for?
Visual inspection zone. It is the ICAO Doc 9303 term for the part of an identity document intended to be read by a person rather than a machine.
Is the portrait part of the VIZ?
Yes. The printed portrait sits in the visual inspection zone, which is one of the main reasons the VIZ has to be captured — it is the reference image for biometric matching against a live selfie.
Why is reading the VIZ harder than reading the MRZ?
The MRZ has fixed character positions, a fixed typeface, and an identical structure worldwide. The VIZ varies in layout, typeface, language, and background design across every issuing authority and every design generation, so it needs document-specific layout models rather than positional parsing.
Do all identity documents have a VIZ?
Effectively yes — any document meant to be read by a person has one. What varies is whether it also has an MRZ. Most US driver licenses, for example, have a VIZ and a barcode but no machine-readable zone.
