부산대학교 도서관

The proliferation of deep learning in medical image processing has spurred the development of a diverse range of medical imaging applications, leading to exponential growth in treatment and diagnostic solutions for various medical images issues. In the age of the Internet of Things (IoT), establishing high levels of privacy and security for medical data in Medical Cyber-Physical Systems (MCPS) is critical to facilitating the progress of complicated medical imaging diagnostic applications. The authors' goal in this research is to create and analyze a secure hospital management system geared toward MCPS for brain tumor identification. The proposed system is designed in three distinct phases and then all of them are integrated to complete the design. In the first phase, a smart healthcare system is introduced to deliver effective health services, particularly targeted towards patients with brain tumors. To achieve this, an application is developed, compatible Microsoft-based operating systems. This application enables patients to access the system either in person or remotely. As a result, patient data remains secure, accessible only to the hospital and the patient. The application involves uploading the patient's MRI images, followed by entering a unique 10-digit number to access the predicted results. In the second phase, the authors propose a deep learning-based tumor detection method specifically designed for brain MRI scans. Lastly, AES-128 algorithm is integrated with the proposed platform for secured medical image storage on the server and the transmission through internet from client to server and back to client after prediction. The proposed system has achieved a level of performance that competes with state-of-the-art (SOTA) methods, as demonstrated through various deep learning classifiers.