Novel Aspects of Secure Medical Image Encryption using Block-Chain and Transformer-Based Deep Learning Model

The vulnerability of medical images is a significant challenge in state-of-the-art telemedicine and Internet of Things (IoT)-based healthcare systems. This paper aims to develop a more comprehensive framework for secure encryption of medical images that fuses blockchain technology into a Transformer-based deep learning model (EB-TDLM) to strengthen data security, privacy, and integrity. And because blockchain is decentralized, data can be securely and immutably contained — meaning less compromise with unauthorized access and data breaches. Moreover, the EB-TDLM model adopts Transformer-driven key generation and upgraded encryption methods, enhancing the encryption robustness greatly. We thoroughly evaluate the proposed framework with experimental analysis and show its effective performance on multiple security metrics. Model has obtained an entropy value of 16.99, which indicates that there is high randomness present in encrypted images. It is highly sensitive to pixel change, achieving 99.99% NPCR (Number of Pixels Change Rate) and UACI (Unified Average Changing Intensity) of 39.49%, and showing a high resistance to differential attack. Moreover, the encryption has very low MSE (0.05), which guarantees almost lossless reconstruction of medical images. Structural integrity, reflected by Structural Similarity Index (SSIM) of 1.0, is crucially maintained in reconstructed images. These findings position the EB-TDLM framework as an exceptionally reliable and efficient approach to safeguarding medical data in IoT-enabled healthcare environments. The proposed system serves as a scalable and effective mechanism for further securing sensitive medical images in contemporary healthcare networks by overcoming key challenges in telemedicine security.