What is the difference between the MRZ and the visual zone?

The MRZ and the visual zone are the two main text areas on an identity document, and the difference comes down to who they were designed for and how they behave as data sources.

The MRZ is built for machines. It uses a fixed, ICAO-standardized layout, a specific monospaced font chosen for easy automated recognition, and a restricted character set of capital letters, digits, and filler characters.

Because the format is globally standardized, software can extract data from an MRZ without needing to know in advance exactly which country or document type it's looking at. The field positions are the same everywhere.

The MRZ also includes check digits, letting software independently verify that the data hasn't been altered.

The visual zone is built for people. It's the printed name, address, photo, signature, and other details laid out however the issuing country or authority chose to design their document, using whatever fonts, languages, and layouts they see fit.

There's no global standard governing this layout the way there is for the MRZ, which means extracting data from it reliably requires software to recognize the specific template for each document type and country, of which there are many thousands in circulation globally.

In terms of content, the two overlap significantly but aren't identical. Both typically include the name, date of birth, and document number, exactly what makes cross-referencing them useful for fraud detection.

But the visual zone usually includes fields the MRZ doesn't, like a full address, and obviously includes the photo and signature, which have no equivalent in the MRZ at all.

For accuracy, MRZ extraction tends to have an edge precisely because of its standardization, while visual zone extraction, despite being technically harder, provides broader field coverage and the photo needed for biometric matching.

Because the two sources have different strengths, ScanDoc extracts and cross-validates data from both rather than relying on just one, using the MRZ's speed and check-digit validation alongside the visual zone's broader field coverage and photo capture.