Deep Ensemble of VGG, ResNet and Inception for Multimodal Authentication System
Latha Krishnamoorthy, Ammasandra Sadashivaiah Raju · 2024
Biometrics have been widely employed in security systems worldwide over the past few decades. However, each unimodal biometric method has its limitations and vulnerabilities, such as fingerprint systems being susceptible to spoof attacks. For overcoming such limitations, designing a multimodal biometric system that combines multiple biometric modalities would enhance performance and bolster security against spoof attacks. This article presents a secure multimodal biometric authentication model by integrating electrocardiogram (ECG) and lip images using a deep transfer learning approach. The traditional methods suffer from authentication accuracy related issues therefore proposed approach considers a complete approach which includes data pre-processing, combined feature extraction and ensemble model of pre-trained deep transfer learning models to improve the authentication accuracy. The experimental analysis shows that the proposed approach has reported the overall accuracy as 98%.