Fingerprint-based authentication systems are widely used across various domains because of their reliability and uniqueness. However, storing fingerprint minutiae as templates in a database poses significant security and privacy risks. Various studies indicate that fingerprints can be reconstructed from stored templates, which underscores the need for stronger template protection schemes. Cancelable fingerprint templates are among the best solutions to ensure that no original information is disclosed if a template is compromised. This study introduces a novel Tetrahedron-based Transformation Technique (TBTT) designed to generate irreversible and cancelable fingerprint templates. The method applies a geometric transformation to divide the plotted minutiae into three triangular parts, after which each segment undergoes a separate rotation in a defined order. The proposed approach produces a robust cancelable template that is resistant to inversion and reconstruction, while addressing risks associated with cross-matching, linking, and template inversion. Comprehensive experiments were conducted on benchmark datasets from the Fingerprint Verification Competition 2004 (FVC2004), including all accessible fingerprint images. The average recognition accuracy for datasets DB1 to DB4 is 98.86%, 95.23%, 96.17%, and 94.73%, respectively, with an average recognition time of 0.24 ms. The results demonstrate that the proposed method provides high matching accuracy and significantly improves template security.
Yoogesh et al. (Mon,) studied this question.