What Is It?
The invention for “Enhanced Comparison of Fingerprint Data” significantly increases recognition accuracy, when comparing fingerprints. It can be used for better identification and privacy protection. To ensure this maximum interoperability among vendors, the invention complies with the widely-adopted international ISO or ANSI standards. Thus, the invention can be easily implemented in most commercially available products.
Fingerprints are unique, very detailed, and difficult to change. They are therefore ideally suited to identify a person and with this protect private data from theft and misuse. Fingerprint identification is an IT based process of comparing two instances of friction ridge skin impressions (minutiae) from fingers to determine whether these impressions are from the same individual. Today there is no 100% accuracy when comparing fingerprints. With the invention for “Enhanced Comparison of Fingerprint Data” the recognition accuracy will be significantly increased. It offers the best trade-off between false matches (algorithm indicates fingerprints are the same, while not being the same) and false non-matches (algorithm indicates fingerprints are not the same while being the same).
For this the invention uses a novel data representation for fingerprint minutiae data, a so-called two-layer mega-vicinity structure. By comparing fingerprint minutiae data, which are transformed according to the present invention, a much higher recognition accuracy is achieved than when comparing original minutiae data directly.
Use Cases and Markets
“Enhanced Comparison of Fingerprint Data” can be used for all fingerprint recognition based applications.
Whenever you need to protect data from theft and misuse; examples:
- Personal device access control – e.g. when accessing your smart phone
- Personal digital assets management – e.g. when accessing your files or exchange of data with third parties
- Online identity authentication – e.g. when accessing your bank account
Whenever you need to protect society from potential harm; examples:
- Critical infrastructure access control – e.g. in the energy, ITC, transport and financial services sector
- Border control system – e.g. for faster and more accurate immigration processes
- Law enforcement use – e.g. convict offenders based on unambiguous fingerprints
The invention is relevant for providers of fingerprint recognition technology and system solutions as well as industrial users with identity management solutions. The fingerprint biometrics market in total is expected to grow to about $12 Billion by 2020, with government (law enforcement, border control, eGovernment projects) and enterprise (smartphone security, mobile payment, physical and logical access control) sectors driving growth (source: Global Industry Analysts, April 2015).
The Combination of “Enhanced Comparison of Fingerprint Data” and "Biometrics and Secret Key Binding Scheme"
To increase data protection, the present invention can be combined with the "Biometrics and Secret Key Binding Scheme", another invention which was successfully funded through CrowdPatent. Tests have confirmed that also the encrypted data representation can be compared, and that no performance degradation happens. Consequently, the present invention provides for an enhanced comparison of fingerprint data, for an increased recognition performance, and for better data protection if combined with the first crowdfunded invention. In particular, the data format is preserved after protection so that there is no need to change the comparator used for those plain transformed representations.
“Our invention enables the significantly increased recognition accuracy, when comparing fingerprints. This will lead to a better protection of personal data from theft and misuse. As our invention is compatible to ISO and ANSI standards – the widely-adopted international standards – it can be used on almost all commercially available products.
In short: the invention provides the best combination of interoperability, recognition performance, and security.”
Performance testing results
The invention has been tested in recognition performance on the public fingerprint database FVC2002DB2 and FVC2006DB2 and shows high performance in terms of low error rates as shown in the following table and graphs. Note that thanks to the fact that the invention preserves the minutiae data structure well even after the protection, the recognition performance achieved by the invention has still a big potential to improve by a third party's minutiae comparison algorithm.
The inventors offer the opportunity of trial use of an academic version of the implemented invention, i.e., an ISO minutiae template comparator with- or without template protection functionality starting from August 2015. If you are interested in a trial use of the invention, send your request to email@example.com for a registration and agreement signing.