To reduce the attack vulnerability of automated biometric recognition systems based on the recognition of the fingerprint characteristic features, it required the design and implementation of a new fingerprint-scanning technology. The new technology could be implemented over time based on optical coherence tomography.
An essential advantage of this new technology is the possibility of acquiring three- dimensional images of the fingerprint characteristic feature. This produces, built on the coherence tomography fingerprint reader as a result of detection, multi-layered or three-dimensional fingerprint samples.
With the introduction of a new class of fingerprint scanners and therefore also a new class of fingerprint samples, new algorithms must be created for processing the new class of fingerprint samples.
The aim of this preparation is the formation and realisation of a new quality algorithm for the estimation of the new class of fingerprint samples. The new quality algorithm was realised based on the Tensor Voting Framework. Its performance was evaluated on a data set consisting out of multi-layer fingerprint samples. The result of the evaluation did show that, the performance of the new quaulity algorithm is outperformed by most of the state of the art quality algorithms.