Scholarlink

As numerous sectors undergo digitization, they generate substantial amounts of data, making the effective security of this data an urgent concern. In recent years, biometrics has undergone significant maturation, with many devices employing fingerprint, facial, and even hand recognition seamlessly integrated into information security systems. This paper presents an advanced biometric recognition system utilizing Finger-Knuckle-Print (FKP) features, employing a single-hidden layer feed-forward neural network known as the Extreme Learning Machine (ELM). The proposed method offers high generalization capability while ensuring rapid learning efficiency. For feature extraction, we integrate Local Phase Quantization (LPQ) and Histogram of Oriented Gradients (HOG) techniques to enhance recognition performance. A series of experiments conducted with a well-known and available FKP dataset, including four fingers of each person out of 165 persons, demonstrated the effectiveness of this scheme, especially when employing the improved version of ELM called Rough ELM (RELM), which outperforms many existing methods in the literature.

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

Robust Biometric Systems: A Secure Framework Leveraging Extreme Learning Machine

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yacine Belhocine,
  • Abdallah Meraoumia,
  • Hakim Bendjenna,
  • Mohammed Saigaa

摘要

As numerous sectors undergo digitization, they generate substantial amounts of data, making the effective security of this data an urgent concern. In recent years, biometrics has undergone significant maturation, with many devices employing fingerprint, facial, and even hand recognition seamlessly integrated into information security systems. This paper presents an advanced biometric recognition system utilizing Finger-Knuckle-Print (FKP) features, employing a single-hidden layer feed-forward neural network known as the Extreme Learning Machine (ELM). The proposed method offers high generalization capability while ensuring rapid learning efficiency. For feature extraction, we integrate Local Phase Quantization (LPQ) and Histogram of Oriented Gradients (HOG) techniques to enhance recognition performance. A series of experiments conducted with a well-known and available FKP dataset, including four fingers of each person out of 165 persons, demonstrated the effectiveness of this scheme, especially when employing the improved version of ELM called Rough ELM (RELM), which outperforms many existing methods in the literature.