Scholarlink

<p>Cancelable biometric systems aim to protect user privacy by transforming biometric data into secure, non-invertible representations. This work proposes a cancelable multimodal biometric verification system based on a nonlinear fusion function and chaos-based encryption. The system utilizes a Modified Inverse Tangent Function (MITF) to perform a fusion of two biometric traits, face and iris. To ensure cancelability and resistance to inversion attacks, the fused template is encrypted using a novel confusion-diffusion architecture driven by a logistic map. Verification is achieved by computing the correlation between stored and query templates in their transformed cancelable forms. Experiments conducted on the ORL and CASIA-Iris databases show that the proposed system achieves perfect verification performance, with an Equal Error Rate (EER) of zero and an Area Under the ROC Curve (AROC) of one. The templates exhibit high entropy exceeding 7.95 and minimal visual similarity to the fused images, validating their non-invertibility. These results demonstrate that the proposed scheme effectively combines robustness, privacy protection, and computational efficiency, positioning it as a promising solution for high-security applications requiring robust and efficient biometric data protection.</p>

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

Cancelable multimodal biometric system through inverse tangent function and chaotic map

  • 期刊论文

  • 发表日期：

    2026年2月9日

查看全文
  • Adil Badreddine,
  • Naceur-Eddine Boukezzoula,
  • Tewfik Bekkouche

摘要

Cancelable biometric systems aim to protect user privacy by transforming biometric data into secure, non-invertible representations. This work proposes a cancelable multimodal biometric verification system based on a nonlinear fusion function and chaos-based encryption. The system utilizes a Modified Inverse Tangent Function (MITF) to perform a fusion of two biometric traits, face and iris. To ensure cancelability and resistance to inversion attacks, the fused template is encrypted using a novel confusion-diffusion architecture driven by a logistic map. Verification is achieved by computing the correlation between stored and query templates in their transformed cancelable forms. Experiments conducted on the ORL and CASIA-Iris databases show that the proposed system achieves perfect verification performance, with an Equal Error Rate (EER) of zero and an Area Under the ROC Curve (AROC) of one. The templates exhibit high entropy exceeding 7.95 and minimal visual similarity to the fused images, validating their non-invertibility. These results demonstrate that the proposed scheme effectively combines robustness, privacy protection, and computational efficiency, positioning it as a promising solution for high-security applications requiring robust and efficient biometric data protection.