Scholarlink

<p>One of the main challenges in fingerprint recognition lies in assessing and ensuring the quality of fingerprint images used in biometric systems. In this paper, we address this challenge by focusing on the construction of a high-quality fingerprint database and leveraging it to validate a novel quality assessment approach. Our primary contribution lies in the development of a comprehensive database explicitly labeled with ground truth on fingerprint image quality, ensuring reliable data. This database serves as a benchmark resource for the biometric community, facilitating standardized evaluation and comparison. Building upon this dataset, we introduce an innovative two-level possibilistic modeling framework for fingerprint image quality assessment. At the first level, local quality indicators are modeled using possibility distributions, effectively capturing uncertainties in image quality assessment. The second level aggregates these local indicators into a global quality score, providing a holistic and flexible representation of fingerprint image quality. To further enhance the robustness of our approach, we introduce discriminative quality indicators that capture subtle variations in fingerprint image quality and employ a sensitivity-aware feature selection process. The effectiveness of our framework is rigorously validated using our newly constructed database, demonstrating its capability to model intricate quality variations across different fingerprint samples. Unlike traditional methods, our approach is database-agnostic, requiring no parameter tuning, thus offering a more adaptable and accurate solution for fingerprint quality assessment. By providing a fingerprint image dataset explicitly labeled by quality, this work establishes a solid foundation for future advancements in fingerprint image quality analysis and biometric applications. Please refer to <a href="https://github.com/Sonda09/FIQM.git">https://github.com/Sonda09/FIQM.git</a>, where the constructed database, the LQIs and PQIs measurements are given.</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

Exploring possibilistic fingerprint image quality analysis: a soft computing approach for biometric database management

  • 期刊论文

  • 发表日期：

    2025年3月11日

查看全文
  • Sonda Ammar Bouhamed

摘要

One of the main challenges in fingerprint recognition lies in assessing and ensuring the quality of fingerprint images used in biometric systems. In this paper, we address this challenge by focusing on the construction of a high-quality fingerprint database and leveraging it to validate a novel quality assessment approach. Our primary contribution lies in the development of a comprehensive database explicitly labeled with ground truth on fingerprint image quality, ensuring reliable data. This database serves as a benchmark resource for the biometric community, facilitating standardized evaluation and comparison. Building upon this dataset, we introduce an innovative two-level possibilistic modeling framework for fingerprint image quality assessment. At the first level, local quality indicators are modeled using possibility distributions, effectively capturing uncertainties in image quality assessment. The second level aggregates these local indicators into a global quality score, providing a holistic and flexible representation of fingerprint image quality. To further enhance the robustness of our approach, we introduce discriminative quality indicators that capture subtle variations in fingerprint image quality and employ a sensitivity-aware feature selection process. The effectiveness of our framework is rigorously validated using our newly constructed database, demonstrating its capability to model intricate quality variations across different fingerprint samples. Unlike traditional methods, our approach is database-agnostic, requiring no parameter tuning, thus offering a more adaptable and accurate solution for fingerprint quality assessment. By providing a fingerprint image dataset explicitly labeled by quality, this work establishes a solid foundation for future advancements in fingerprint image quality analysis and biometric applications. Please refer to https://github.com/Sonda09/FIQM.git, where the constructed database, the LQIs and PQIs measurements are given.