Title: Preventing playback attacks in fingerprint biometrics through four-level hashing
Authors: H. Mohamed Khan
Addresses: Department of Computer Science, Noorul Islam College of Arts and Science, Kumaracoil, Thuckalay, Kanyakumari District, Tamil Nadu, India
Abstract: Replay (playback) attacks threaten fingerprint biometric systems by re-using previously captured data to gain unauthorised access. This paper proposes fingerprint matching using four-level hashing method (FMFLHM) - a layered approach combining timestamp transformation, fingerprint trait modification, XOR mixing, triangular hashing (three iterative hashes), and AES encryption. FMFLHM fuses ridge orientation and Gabor-HOG descriptors, binds them to time and a server nonce, and transmits an AES-encrypted triangular hash for server verification. Experiments on NIST-302 and MOLFDB datasets show FMFLHM reduces false acceptance and rejection rates and gives higher authentication accuracy (98.09% on NIST-302). We discuss key-management, scalability for resource-constrained devices, and limitations versus advanced spoofing techniques. FMFLHM strengthens resistance to replay attacks and offers a practical route to more robust fingerprint authentication.
Keywords: NIST; playback attack; hashing; cybersecurity; ridge orientation; histogram of gradients.
DOI: 10.1504/IJICS.2026.151306
International Journal of Information and Computer Security, 2026 Vol.29 No.2, pp.113 - 134
Accepted: 14 Jun 2025
Published online: 22 Jan 2026 *