Hand biometrics experts discuss synthetic data, image quality standards

Hand biometrics are like your favorite character actor: omnipresent, underappreciated, and typically outshined by the shinier faces in the room. The European Association of Biometrics’ Council of Wisdom gathered this week to give some love to hand biometrics in a recent webinar.

The discussion sees experts field questions from audience members about a variety of topics related to all aspects of hand biometrics, including fingerprint, palm print and palm vein modalities.

The first concerns the maturity of synthetic fingerprint generation tech, and to what extent it can be trusted. Martin Drahansky, from the Police Academy of the Czech Republic in Prague, says the best algorithms are now up to the task. “The trouble is to generate synthetic fingerprints which have real changes of the skin structure,” including things like injuries, skin diseases or other environmental factors. So, while the fingerprints themselves are high quality, they don’t change like real skin. Generative AI is helping to address this, but is not especially effective yet.

Matteo Ferrara of the University of Bologna says “synthetic data in biometrics is very important” because of privacy issues, and the difficulty of collecting real data. But, he emphasizes the importance of mixing synthetic data and real data for training algorithms, and underscores the importance of testing published results on real data.

A further question concerns standards around image quality. They exist for fingerprint and face, but not for palm vein biometrics. Raymond Veldhuis, a professor at the University of Twente, says “there is a great need for quality standards for vein pattern recognition.”

“But as far as I know, they don’t exist.”

The question of how to protect data is raised, specifically in the context of the GDPR and cross-jurisdictional sharing. On this, there are no concrete answers; all participants acknowledge the complexity of sharing data responsibly and in compliance with stringent data protection laws.

The maturity of contactless fingerprint recognition for smartphone cameras comes up; Veldhuis says the tech is “not yet mature,” and questions “whether or not there would be a real market for this.” Drahansky suggests that there could be applications for police forces in the U.S. – and also hints at companies developing fingerprint ID for bank account opening. “Maybe it will be the future,” he says. “But it’s something we are not using at the moment.”

A final question concerns whether palm prints are a better option than face biometrics for purposes of consent, as defined by GDPR. But Drahansky says it’s too difficult to capture a high quality palm print to make it practically feasible in most use cases.

In the end, hand biometrics are likely to remain in the shadow of facial matching – for now. Gestural biometrics are being pitched as a potential solution to age assurance for adult content. And plenty of people still unlock their smartphone with their fingerprint.

The overarching message might be summed up as follows: with all the different use cases, there’s room enough in the world for every biometric.

Article Topics

biometrics research  |  EAB  |  EAB 2025  |  European Association for Biometrics  |  fingerprint biometrics  |  hand biometrics  |  palm biometrics  |  palm vein biometrics  |  synthetic data