Match-On-Card

Match-On-Card -- On-Chip Biometric Verification

Match-On-CardMatch-On-CardBiometricBiometric matching performed inside the smart card chip.Click to view → (MOC) is a biometric authentication approach where the comparison between a live biometric sample and the stored reference template occurs entirely within the smart card's secure element, rather than on an external host system. This architecture ensures that sensitive biometric data never leaves the card, providing maximum privacy protection and eliminating network-based attack vectors.

How MOC Works

During biometric enrollment, a reference template (typically fingerprint minutiae data) is extracted from the cardholder's biometric sample and stored in the card's EEPROM. At verification time, a reader-connected sensor captures a fresh fingerprint image, extracts a candidate template, and sends it to the card via an APDU command. The card's on-chip processor executes the matching algorithm -- comparing minutiae points, ridge patterns, and spatial relationships -- and returns only a binary match/no-match result. The original reference template is never exposed to the external system.

Performance and Accuracy

Modern smart card chips with dedicated crypto coprocessors can execute a MOC comparison in 1-3 seconds, depending on the template format and matching algorithm complexity. Accuracy metrics follow the FAR/FRR framework: typical MOC implementations achieve a False Accept Rate below 0.01% with a False Reject Rate under 3%. Template sizes range from 300 to 800 bytes for fingerprint minutiae (ISO 19794-2), fitting comfortably within the card's memory constraints.

Applications

MOC is deployed in biometric payment cards where it replaces PIN entry for cardholder verification in EMV transactions. Government ID programs use MOC on PIV and CAC cards for high-assurance physical access control. ePassport gates at border crossings can leverage MOC when the travel document supports on-card matching, though most current deployments still use external matching. The CBEFF framework ensures that biometric templates stored on cards conform to an interoperable data format regardless of the card platform (JavaCard or MULTOS).