Scholarlink

The security offered by biometric technology plays a crucial role in its reliability and acceptance, which is hampered by the leakage of biometric templates and undermines the user’s privacy. This eventually causes permanent identity loss for a user. To overcome this issue, we propose a robust template protection technique for fingerprint user templates based on a Fuzzy Commitment Scheme (FCS). The original template is binarized by utilizing a modification of Binarized Phase Spectrum (BiPS), which is a binary string of fixed-length extracted by quantization of the phase spectrum computed from the minutiae set. Further, the generated binary template is secured by utilizing the FCS. To align the fingerprint images, a Principal Component Analysis (PCA) based technique has been utilized. The proposed technique achieves the FRR of 16.05%, 9.1%, and 4% on FVC2002 DB1 employing FVC protocol, enrollment of two images, and 1-vs-1 protocol, respectively, at zero FAR.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Secure Fingerprint Authentication Using a Robust and Effective Biometric Cryptosystem

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Vivek Singh Baghel,
  • Surya Prakash,
  • Ashish Banala,
  • Aravind Ravikumar

摘要

The security offered by biometric technology plays a crucial role in its reliability and acceptance, which is hampered by the leakage of biometric templates and undermines the user’s privacy. This eventually causes permanent identity loss for a user. To overcome this issue, we propose a robust template protection technique for fingerprint user templates based on a Fuzzy Commitment Scheme (FCS). The original template is binarized by utilizing a modification of Binarized Phase Spectrum (BiPS), which is a binary string of fixed-length extracted by quantization of the phase spectrum computed from the minutiae set. Further, the generated binary template is secured by utilizing the FCS. To align the fingerprint images, a Principal Component Analysis (PCA) based technique has been utilized. The proposed technique achieves the FRR of 16.05%, 9.1%, and 4% on FVC2002 DB1 employing FVC protocol, enrollment of two images, and 1-vs-1 protocol, respectively, at zero FAR.