IMPROVING HUMAN RECOGNITION IN BIOMETRIC SYSTEMS USING NEGATIVE SELECTION ALGORITHM (NSA)
Loading...
Date
2023-11-04
Journal Title
Journal ISSN
Volume Title
Publisher
ANNALS of Faculty Engineering Hunedoara – INTERNATIONAL JOURNAL OF ENGINEERING
Abstract
The Negative Selection Algorithm (NSA), inspired by the human immune system, has wide–ranging applications including intrusion detection,
anomaly detection, and pattern recognition, but its application in human recognition has not been thoroughly explored. This study investigates its potential
for human identification, particularly in bi–modal systems that combine physiological and behavioral traits. 2400 sample images from 200 individuals were
collected and divided into training, testing, and validation data sets. Images were pre–processing and principal component analysis was used to select salient
features. These selected features were fused at the feature level using the weighted average method and NSA was used as classifier. The behavioral feature–
based system achieved a remarkable 95% accuracy rate, with true positive (TP) and true negative (TN) rates of 141% and 144%, respectively. In comparison,
the physiological traits–based system achieved an 89% accuracy rate. The voice–based uni–modal system outperformed others, with TP and TN rates of
131% and 134%, respectively, with accuracy rate of 88.33%. These findings established the advantages of combining biometric features to enhance system
accuracy. It also demonstrates that NSA can significantly improve the precision of biometric systems classification. The developed biometric systems can be
emulated in any system that requires ultra–level of security.
Description
Keywords
Citation
Balogun, M. O., Jimada-Ojuolape, B., & Odeniyi, A. L. (2024). Enhancing Automated Face Recognition with Makeup Detection: A CNN-Based Approach. COVENANT JOURNAL OF ENGINEERING TECHNOLOGY, 8 (1), 1-10.