Reliable and privacy-preserving multi-instance iris verification using Paillier homomorphic encryption and one-digit checksum

• Published in: Signal, image and video processing Vol. 18; no. 4; pp. 3723 - 3735
• Main Authors: Morampudi, Mahesh Kumar, Gonthina, Nagamani, Bojjagani, Sriramulu, Sharma, Neeraj Kumar, Veeraiah, Duggineni
• Format: Journal Article
• Language: English
• Published: London Springer London 01-06-2024; Springer Nature B.V
• Subjects: Authentication; Biometrics; Cloud computing; Computer Imaging; Computer Science; Image Processing and Computer Vision; Multimedia Information Systems; Original Paper; Pattern Recognition and Graphics; Privacy; Signal,Image and Speech Processing; Verification; Vision; Privacy-preserving; Homomorphic encryption; Biometric authentication system; Multi-instance iris
• ISSN: 1863-1703; 1863-1711
• DOI: 10.1007/s11760-024-03036-0

The utilization of a biometric authentication system (BAS) for reliable automatic human recognition has increased exponentially in recent years over traditional authentication systems. Since the biometric traits are irrevocable, two important issues such as security and privacy still need to be addressed in BAS. Researchers explore homomorphic encryption (HE) to propose several privacy-preserving BAS. However, the correctness of the evaluated results computed by the cloud server on the protected templates is still an open research challenge. These methods are able to conserve the privacy of biometric templates but unable to check the correctness of computed result results in false reject or accept. To overcome this issue, we suggest a reliable and privacy-preserving verifiable multi-instance iris verification system using Paillier HE and one-digit checksum (PVMIAPO). Modified local random projection is implemented on the fused iris template to produce the reduced template. Later, Paillier HE is applied on the reduced template to create the protected template. The result returned by the third party server is verified using the one-digit checksum. The efficiency of PVMIAPO is verified by experimenting with it on SDUMLA-HMT, IITD, and CASIA-V3-Interval iris databases. PVMIAPO gratifies the irreversibility, diversity, and revocability properties. PVMIAPO also obtains fair performance in contrast to the existing methods.