Biometrics are essential in facility access control where human characteristics are used for person identification. Fingerprint is a basic identity for verification. Fingerprint segmentation is an open problem in biometrics, hence it is the first process towards the utilization of fingerprints for human recognition because each individual has a unique and permanent fingerprint sample. Considering this uniqueness, Fingerprint-based human verification is used in several applications for a an era. A fingerprint image is considered as a pattern which composed of two regions, foreground and background. The foreground contains the significant information utilized in the fingerprint identification systems. However, the background is usually noisy, distorted and unstable region that plays an important role in the extraction of false minutiae in the fingerprint system. In order to avoid this, we use fingerprint segmentation to isolate foreground/background fingerprint composites. Threshold, Entropy and Type-2 fuzzy-logic based fingerprint segmentation are presented and compared in this work. furthermore a demonstration of Fingerprint-based human verification approach that employee the presented segmentation techniques is presented and evaluated using ROC curve to prove the validity and reliability of proposed methodology.
Digital Object Identifier (DOI)
S. Khalifa, Hany; I. Wahhab, H.; N. Alanssari, A.; and A. O Ahmed Khfagy, M.
"Fingerprint Segmentation Approach for Human Identification,"
Applied Mathematics & Information Sciences: Vol. 13:
4, Article 1.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol13/iss4/1