AI-Driven Multimodal Authentication: Strengthening Healthcare Biometric Security with GRNs
M Vijay, Bandi Praveen Kumar, Sorakayala Vishnu Vardhan, V. Siva Reddy · 2025
In this study we propose a new multimodal biometric authentication model based on gated recurrent networks (GRNs) aimed at improving the security of biometric data collected from existing healthcare providers. Authentication starts by collecting biometric data, such as fingerprints, ECG signals, and PPG signals. The biometric data we collected is adjusted through multiple pre-processing steps This includes eliminating unwanted noise, artefacts, and inconsistencies with the iCanClean algorithm, which has demonstrated efficacy in eliminating artefacts correlated with noise electrodes. This leads the study to propose the use of a Squeeze-and-Excitation Network (SE-Net) attention framework along with pattern-uneven wavelet functions to perform feature extraction in an effective manner. To solve longer encoding sequence problems, such as long-range dependency and gradient vanish, this research proposes using a Gated Recurrent Unit (GRU) to perform the authentication decision. Based on the number of training subjects, the proposed GRU accuracy value of 99.24% displays good performance. Acknowledgements: We believe that this multimodal authentication model based on GRNs for biometrics data protection in healthcare will have great potential for future work.