High Impact Factor : 4.396 icon | Submit Manuscript Online icon | UGC Approved icon

Statistical Features Based Gender Identification Using SVM


Shivanand S Gornale , Rani Channamma University, Belagavi; Abhijit Patil , Rani Channamma University, Belagavi; Prabha, KASC College, Bidar


Gender Identification, Statistical Features, Fingerprints, SVM, QDA, LDA


The gender identification based on any biometrics is essential to build applications such as human-computer interaction, content-based indexing, searching, targeted advertising, surveillance, biometrics, and demographic studies. The Fingerprints are one of the most notable biometric technologies and are considered as legitimate evidence all over the world to identify a person. The Fingerprint traits can be used effectively for the estimation of gender information. This paper presents a technique for gender identification based on simple statistical properties of the fingerprints. These are extracted from 740 Fingerprint images. A set of 15 statistical properties of a Fingerprint image is used to form a feature vector. Then, the traditional Support Vector Machine (SVM) is employed to draw the decision using the feature metrics of size 740x15. The experimental results stood at 87% and 89.15% as encouraging evidence.

Other Details

Paper ID: NCACCP047
Published in: Conference 8 : NCACC-2016
Publication Date: 01/10/2016
Page(s): 241-244

Article Preview

Download Article