What is a barcode/2D barcode extraction (PDF417, QR, Aztec), and when is it used?

Barcode extraction is the process of reading and decoding machine-readable barcodes printed on identity documents, distinct from both MRZ reading (which processes standardized OCR text) and visual zone OCR (which reads the human-readable printed text). A barcode encodes data in a visual pattern, parallel lines for 1D barcodes, or a grid of dots and shapes for 2D formats like PDF417, QR, and Aztec, that a scanner decodes directly into structured data, without needing to "read" text the way OCR does.

The most common context for barcode extraction in identity documents is North American driver's licenses and ID cards, which typically carry a PDF417 barcode on the back encoding data according to the AAMVA (American Association of Motor Vehicle Administrators) standard: name, date of birth, license number, address, expiry date, and vehicle class information, among other fields. Because these documents generally don't include an ICAO-format MRZ, barcode decoding is often the primary structured data source for this document category, rather than a secondary check alongside MRZ data.

QR codes and Aztec codes show up in various other document contexts, including some boarding passes (following IATA's barcoded boarding pass standard), certain vaccination or health certificates, and some newer digital or hybrid identity documents. These formats can encode more data in a smaller physical space than older 1D barcodes, part of why they've become more common as document designers look to pack in more structured, verifiable data without expanding the barcode's footprint.

Barcode decoding is generally very reliable compared to OCR, since it's reading a structured visual pattern rather than interpreting variable typography, though it depends on the barcode itself being undamaged and captured clearly. A scratched or partially obscured barcode can still fail to decode correctly.

ScanDoc's document scanning includes barcode reading alongside MRZ and visual zone OCR, decoding PDF417, QR, and other common 2D barcode formats so documents like North American driver's licenses, which rely primarily on barcodes rather than an MRZ, are still fully processed as part of the standard scanning flow.