Comparison results of similar schemes.

Most existing secure biometric authentication schemes are server-centric, and users must fully trust the server to store, process, and manage their biometric data. As a result, users’ biometric data could be leaked by outside attackers or the service provider itself. This paper first constructs the EDZKP protocol based on the inner product, which proves whether the secret value is the Euclidean distance of the secret vectors. Then, combined with the Cuproof protocol, we propose a novel user-centric biometric authentication scheme called BAZKP. In this scheme, all the biometric data remain encrypted during authentication phase, so the server will never see them directly. Meanwhile, the server can determine whether the Euclidean distance of two secret vectors is within a pre-defined threshold by calculation. Security analysis shows BAZKP satisfies completeness, soundness, and zero-knowledge. Based on BAZKP, we propose a privacy-preserving biometric authentication system, and its evaluation demonstrates that it provides reliable and secure authentication.

当前主流的安全生物特征认证方案多采用服务器中心化架构，用户需完全信任服务器以存储、处理并管理其生物特征数据。由此一来，用户的生物特征数据存在被外部攻击者或服务提供商自身泄露的风险。本文首先基于内积构造了EDZKP协议，该协议可验证某一秘密值是否对应秘密向量的欧氏距离。随后，结合Cuproof协议，本文提出了一种全新的用户中心化生物特征认证方案BAZKP。在该方案中，所有生物特征数据在认证全程均保持加密状态，服务器无法直接获取明文数据。与此同时，服务器可通过计算判定两个秘密向量的欧氏距离是否处于预先设定的阈值区间内。安全性分析表明，BAZKP满足完备性（Completeness）、合理性（Soundness）与零知识（Zero-knowledge）三大属性。基于BAZKP方案，本文构建了一款隐私保护型生物特征认证系统，经性能评估验证，该系统可提供可靠且安全的认证服务。