Open AccessCross-spectral Gender Classification Using Multi-spectral Face Imaging
Author(s) -
Narayan Vetrekar,
Aparajita Naik,
R. S. Gad
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1921/1/012048
Subject(s) - biometrics , discriminative model , support vector machine , artificial intelligence , pattern recognition (psychology) , facial recognition system , classifier (uml) , computer science , spectral imaging , cross validation , face (sociological concept) , computer vision , geography , remote sensing , social science , sociology
Soft biometrics has been gaining significant attention in biometrics literature for usage in recognition system. Gender classification is one such soft biometric traits that has been studied extensively using visible spectrum range and limited studies beyond visible spectrum. In this paper we present the cross-spectral gender classification using Multi-spectral imaging. We present this study using the facial images captured in nine narrow spectral bands ranging from 530nm to 1000nm. Further, we present an extensive benchmark evaluation based on state-of-the art linear Support Vector Machine (SVM) classifier for facial discriminative features in the cross spectral data in a robust manner. The extensive evaluation results based on 10 fold cross-validation indicates the highest average classification accuracy of 92.84% for cross-spectral study.