Robust Privacy Protection Framework for Medical Images

Online medical care puts patient privacy and medical diagnostic images at risk of being leaked and stolen when transmitted over the network. Therefore, privacy protection of medical images has attracted widespread attention from researchers. Existing privacy protection methods mainly focus on reversible data hiding (RDH) and a small amount of watermarking techniques, but these methods have limited effectiveness in dealing with noise attacks. Therefore, to improve the diagnostic efficiency of doctors while resisting noise attacks during image sharing and ensuring the visual quality of images, we propose a new privacy protection framework for medical images. Initially, we use MedSAM to segment the lesion area in the medical image and then apply contrast enhancement to improve the quality of the medical image. Subsequently, we use blind watermarking technology to embed the patient identity information into the enhanced medical image. Experiments demonstrate that Our method effectively separates and enhances the lesion region while ensuring the hiding of patient identity information, and exhibits high levels of robustness and imperceptibility.