Deep neural network-based secure healthcare framework

Deep neural network-based secure healthcare framework

연구 분야: Cryptography

학회: Neural Computing and Applications

Healthcare stands out as a critical domain profoundly impacted by Internet of Things (IoT) technology, generating vast data from sensing devices as IoT applications expand. Addressing security challenges is paramount for a successful IoT healthcare framework, with blockchain technology offering a decentralized structure for robust data protection and secure data exchange within multi-node IoT networks. The research introduces a secure IoT healthcare diagnostic model empowered by deep neural networks, emphasizing encryption, safe transactions, and healthcare diagnostics as key components. Notably, the model incorporates innovative techniques like the orthogonal particle swarm optimization algorithm for sharing medical images and a neighborhood indexing sequence method for hash value encryption. The development of an optimized deep neural network-based classification model for illnesses, validated through extensive trials, demonstrates superior performance metrics compared to existing decision-making techniques, with significant improvements in f-Measure (96.25%), sensitivity (93.26%), specificity (94.26%), and accuracy (93.26%). This study’s scientific contribution lies in its innovative approach to securing IoT-healthcare diagnosis models, validated performance enhancements using real-world datasets, and insightful recommendations for future research directions, fostering advancements in healthcare technology for enhanced patient care and system efficiency.