Comparative Analysis of Fingerprint-Based Gender Classification Employing Convolutional Neural Network
Lubana Akter, Md. Abdul Based, A.B.M. Toufique-Ul Islam, Elias Ur Rahman · 2024
Gender classification has played a very important role in various domains like criminology, surveillance, human-computer interaction, and commercial purposes. Many previous studies focused on utilizing biometric features like iris, face, gait, and hand shapes for gender classification. This research aims to explore the potential of fingerprints as a feasible direction for gender classification. In this work, Convolutional Neural Network (CNN), K-Nearest Neighbors, & Support Vector Machine are applied on 55,273 fingerprint images to thoroughly assess fingerprint-based gender classification. This system achieved an exceptional accuracy rate of 99.86% with CNN. This outstanding performance highlights the important role that fingerprints may play in accurately recognizing a person's gender by utilizing important characteristics available in the fingerprint data. Thus, this study provides the potential of fingerprints as a reliable biometric identification method for determining gender.