Secure and cancelable fingerprint templates with privacy-preserving mapping and template encoding

Cancelable fingerprint template methods frequently encounter challenges in achieving an optimal balance between security and recognition accuracy, leading to privacy vulnerabilities and reduced system performance. To address these challenges in cancelable fingerprint template protection, we propose a novel approach for secure and cancelable fingerprint templates with privacy-preserving mapping and template encoding. Firstly, we introduce a privacy-preserving feature mapping technique to enhance security and privacy. Secondly, we present Adaptive Momentum Binary Hashing (AMBH), which improves the conversion of binary features, optimizing both feature representation and computational efficiency. Finally, we generate the cancelable secure fingerprint template by applying feature transformations, decimalization, and secure storage in a bloom filter to facilitate efficient storage and conflict management. Experimental evaluations on the FVC2002 and FVC2004 fingerprint databases demonstrate the proposed scheme's superior accuracy, establishing it as a state-of-the-art solution for fingerprint recognition. Furthermore, a comprehensive security analysis confirms that the method adheres to rigorous cancelable biometric template protection standards, ensuring robust irreversibility, unlinkability, revocability, and resistance to various attacks.