Abstract
In today's automated society, authentication has become a necessity. Fingerprint-based authentication is the predominant choice due to its advantages. However, existing biometric methodologies have certain privacy and security challenges as they directly store the extracted minutiae points in the database, potentially enabling malicious entities to reconstruct fingerprints. The proposed approach constructs a Delaunay triangle from minutiae points, and features are extracted using triangle properties (incircle, circumcircle, and excircle). A secure user template is formed by incorporating fingerprint features, generating a secure bitstring, and utilizing a distinct user key achieved through the salting process. The system's security is enhanced by feature mapping. The proposed method demonstrates exceptional revocability, diversity, security, and performance and has been rigorously evaluated on nine databases, namely FVC2000 (DB1 & DB2), FVC2002 (DB1, DB2, DB3, & DB4), FVC2004 (DB1 & DB2), and IIT Kanpur database, which is a large database.
| Original language | British English |
|---|---|
| Article number | 103876 |
| Journal | Computers and Security |
| Volume | 142 |
| DOIs | |
| State | Published - Jul 2024 |
Keywords
- Authentication
- Biometrics
- Delaunay triangle
- Fingerprint
- Privacy
- Security
- Unlinkability
Fingerprint
Dive into the research topics of 'FinTem: A secure and non-invertible technique for fingerprint template protection'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver