Comprehensive Methods for Secure Medical Image Transfer: Encryption and Standardization Techniques

The primary goal of the process is to securely transfer medical images. This involves utilizing MRI datasets for processing and applying resizing and filtering techniques such as Conv2D and Gaussian blur to standardize the images. Additionally, an attack model is developed to enhance information security, starting with a PGD attack to access model weights and followed by a DAG attack to increase loss functions. Encryption is then performed using a Chaos-based encryption model to create a chaotic map and the Henon map for image conversion. A randomly generated key encryption method is utilized to generate a key for secure transfer. The encrypted image is displayed, and in the receiver section, the process is reversed for decryption by inputting the key generated in the encryption model. Finally, the original medical image is retrieved along with parameters such as MSE, MAE, RMSE, and F2 score to assess the accuracy of the recovered image and ensure its correctness.