Regula builds age, gender estimation into ID document scans

Regula is enhancing its document reader software with biometric age and gender estimation based on document photos. The identity verification company says that the new feature will allow its customers to detect identity misuse and attempts to avoid age checks.

Fraud is often carried out with manipulated data on real identity documents, rather than forgeries, the company points out in its announcement. Personal details on IDs like passports or driver’s licenses are altered to allow another individual to use them, to bypass age restrictions or other controls. These types of documents can fool both visual and automated checks.

The new biometric age and gender estimation feature on Regula’s Document Reader SDK analyzes the face printed on the document and estimates a person’s approximate age and gender. It then automatically compares them against the document data, such as date of birth and date of issue, to identify mismatches. The purpose of the new feature is not to determine an exact age or identify an individual but to detect discrepancies and fraud, the company explains.

Regula’s age estimation technology has impressed in evaluations by NIST, topping several categories in the Face Analysis Technology Evaluation (FATE) for biometric facial age estimation (FAE).

“By validating whether biometric attributes derived from document photos align with official document data, we add a new layer to document-based verification – one that focuses on biometric consistency,” says Regula Chief Technology Officer Ihar Kliashchou. “This should help organizations strengthen identity decisions where authenticity checks and visual inspection are no longer sufficient.”

The Document Reader SDK is designed to carry out a series of automated verification checks to confirm the authenticity of identity documents, examining elements such as MRZ data, barcodes, and RFID chips. These checks are supported by the company’s document template database, which encompasses more than 16,000 templates.

Last month, the Latvia-based company also launched a new feature for the Document Reader SDK, which will allow organizations to verify mobile driver’s licenses (mDLs) at scale.

Article Topics

biometric age estimation  |  biometrics  |  face biometrics  |  facial age estimation (FAE)  |  gender recognition  |  Regula