Spoofing detection system for e-health digital twin using EfficientNet Convolution Neural Network

• Published in: Multimedia tools and applications Vol. 81; no. 19; pp. 26873 - 26888
• Main Authors: Garg, Hitendra, Sharma, Bhisham, Shekhar, Shashi, Agarwal, Rohit
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.08.2022; Springer Nature B.V
• Subjects: 1202: Multimedia Tools for Digital Twin; Artificial neural networks; Biometrics; Computer Communication Networks; Computer Science; Cyber-physical systems; Data Structures and Information Theory; Diagnosis; Digital imaging; Digital twins; Electronic devices; Multimedia Information Systems; Neural networks; Special Purpose and Application-Based Systems; Spoofing; Virtual reality; Wearable technology; Spoofing; EfficientNet; Digital twin; e-Health cloud; Authentication system
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-021-11578-5

Digital Twin is the mirror image of any living or non-living objects. Digital Twin and Cyber-physical system (CPS) provides a new era for industries especially in the healthcare sector that keeps track of health data of individuals to provide on-demand, fast and efficient services to the users. In the suggested system, various health parameters of the patients are collected through different health instruments, wearable devices that communicate data to the primary database; used for analysis purposes for better diagnosis and training for automated systems. The primary database in a physical object and parallelly maintain virtual object/digital twin of the same in order of analyzing, summarize and mine data for diagnosis, monitoring the patient in real-time. The e-health cloud data need to be protected from unauthorized access by biometric authentication using iris biometric trait. The proposed paper suggested two phases EfficientNet Convolution Neural Network-based framework for identifying the real or spoofed user sample. The proposed system is trained using EfficientNet Convolution Neural Network on different datasets of spoofed and actual iris biometric samples to discriminate the original and spoofed one.