Secure-by-Disguise: A Systematic Evaluation of Image Disguising for Confidential Medical Image Modeling

A systematic evaluation of image disguising techniques for confidential medical image modeling found that performance varies significantly between tasks, with methods preserving utility for classification but causing substantial degradation in semantic segmentation. The study, which assessed DisguisedNets and NeuraCrypt across four datasets, identified Randomized Multidimensional Transformation as offering the optimal balance of performance and security, while AES-based disguising severely impacted utility.

arXiv:2607.08867v1 Announce Type: new Abstract: Cloud-based deep learning enables large-scale medical image analysis but raises significant privacy concerns when sensitive patient images are outsourced for model development. Image disguising has recently emerged as a promising privacy-enhancing technology PET that transforms images into visually unintelligible representations while preserving information for downstream learning. We established a unified framework to evaluate representative methods, DisguisedNets and NeuraCrypt, across four datasets involving classification and semantic segmentation tasks. Our analysis assessed predictive utility, efficiency, and robustness against reconstruction attacks. Results showed that image disguising performance varies significantly between tasks; while methods preserved utility for medical image classification, they caused substantial degradation in dense semantic segmentation. Specifically, Randomized Multidimensional Transformation RMT offered the optimal balance of performance and security, whereas AES-based disguising severely impacted utility. Furthermore, regression-based reconstruction attacks effective on natural images proved considerably less successful on realistic medical images. These findings provide a systematic assessment of PET suitability for confidential medical AI applications.