BlockDeepNet: A Proposed Framework for the Detection of CT-MRI Imaging Using Blockchain and Deep Learning Architecture

Early diagnosis of brain tumors can control the severity of its spread in the affected area. Numerous deep learning models have been proposed for CT-MRI imaging which act as tool for better data visualization and also automate the diagnosis procedure. Along with that the real-world problem is security while sharing the data among hospitals or remote diagnosis of patients by medical practitioners and researchers. The most important concern in sharing medical data is privacy and to achieve this goal, various collaborative models have been developed to secure data transfer. Through this paper, a proposed collaborative model will be utilized constitutes deep learning model and blockchain mechanism which will collect MRI images from different sources and then training to that data will be provided through a global learning model. By using the mechanism of blockchain technology, the data will be authenticated and in next phase training on globally shared data will be provided via keeping the concern of privacy. Afterward, a deep federated learning model will be implemented for network-based segmentation and classification for multiple classes of the tumor for the improved and secure scenario.