Federated Deep Learning Model for Secured Data Transmission in the Healthcare Sector Using Adaptive and Attention‐Based Residual Capsnet With Blockchain
Kapil Netaji Vhatkar, Prakash Sontakke, Aarti S. Pawar, Ujwal Ramesh Shirode, Dipmala Salunke · Computational Intelligence · 2026
ABSTRACT Existing healthcare applications struggle to process and preserve the huge volume of clinical information. The emerging growth of an “Internet of Things (IoT)” has provided opportunities for medical applications to develop into IoT‐aided models with the ability to handle enormous healthcare data. However, IoT‐aided healthcare systems remain dangerous to different safety and privacy problems, as healthcare data is highly sensitive to loss and tampering caused by unauthorized third‐party access. Embedding blockchain technology into the IoT framework ensures trust across networks and enhances data integrity, privacy, and scalability. Moreover, the increasing number of applications and users creates challenges during the training process, which affects the overall effectiveness and performance of the framework. To address these shortcomings, a blockchain‐enabled federated learning network is proposed. The model enhances the safety and privacy of healthcare information during transmission by combining the strengths of federated learning and blockchain technology. The proposed model uses a Federated Learning‐aided Adaptive and Attention‐based Residual CapsNet (AA‐CapsNet) to securely transmit data over the IoT network. Blockchain confirms that individual official operators may access specific information from an IoT network. This blockchain‐based framework maintains the integrity of sensitive information, making the system more reliable and trustworthy. The consistency of the proposed framework is improved by tuning the attributes of AA‐CapsNet using the Random Parameter Updated Crayfish Optimization (RPU‐CFO) algorithm. Finally, the efficiency of developed arrangement techniques and methodologies is evaluated using different systems of measurement, confirming that a developed method be vigorous and highly dependable.