Scholarlink

<p>A biometric authentication system is used to recognize individual persons based on biological characteristics like veins, gaits, iris, fingerprints, signatures, typing styles, ears, odours, etc. The previous uni-modal biometric authentication systems do not offer improved security and accuracy. So, multi-modal biometric authentication has been introduced in recent studies. However, intra-class variations, computational complexity, scalability, and generalizability are still major problems. Therefore, this study introduces a new Dispersive Flies optimized ResNet (DF-ResNet) with Gaussian Support Vector Machine (GSVM) method for effectively recognizing individuals from the given multi-modal biometrics data like fingerprint, iris, and signature. In this, the DF-ResNet is employed for efficient feature extraction from multi-modalities, and the fused features are given as GSVM to perform the recognition process. The experimental results show that the proposed method attains an improved accuracy of 99.5% in the NIST Special Database and UBIRIS V2 and 97.08% in the CASIA dataset. Thus, the analysis proves the strength of the proposed work over other existing methods.</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

Multi-modal biometric authentication with DF-RESNET and concatenated GSVM

  • 期刊论文

  • 发表日期：

    2025年6月24日

查看全文
  • Shruthi N M,
  • Vinod D S

摘要

A biometric authentication system is used to recognize individual persons based on biological characteristics like veins, gaits, iris, fingerprints, signatures, typing styles, ears, odours, etc. The previous uni-modal biometric authentication systems do not offer improved security and accuracy. So, multi-modal biometric authentication has been introduced in recent studies. However, intra-class variations, computational complexity, scalability, and generalizability are still major problems. Therefore, this study introduces a new Dispersive Flies optimized ResNet (DF-ResNet) with Gaussian Support Vector Machine (GSVM) method for effectively recognizing individuals from the given multi-modal biometrics data like fingerprint, iris, and signature. In this, the DF-ResNet is employed for efficient feature extraction from multi-modalities, and the fused features are given as GSVM to perform the recognition process. The experimental results show that the proposed method attains an improved accuracy of 99.5% in the NIST Special Database and UBIRIS V2 and 97.08% in the CASIA dataset. Thus, the analysis proves the strength of the proposed work over other existing methods.