HideMIA: Hidden Wavelet Mining for Privacy-Enhancing Medical Image Analysis

Xun Lin, Yi Yu, Zitong YU, Ruohan Meng, Jiale Zhou, Ajian Liu, Yizhong Liu, Shuai Wang, Wenzhong Tang, Zhen Lei, Alex Kot

Abstract: Despite the advancements that deep learning has brought to medical image analysis (MIA), protecting the privacy of images remains a challenge. In a client-server MIA framework, especially after deployment, patients' private medical images can be easily captured by attackers from the transmission channel or malicious third-party servers. Previous MIA privacy-enhancing methods, whether based on distortion or homomorphic encryption, expose the fact that the transmitted images are medical images or transform the images into semantic-lacking noise. This tends to alert attackers, thereby falling into a cat-and-mouse game of theft and protection. To address this issue, we propose a covert MIA framework based on deep image hiding, namely HideMIA, which secures medical images by embedding them within natural cover images that are unlikely to raise suspicion. By directly analyzing the hidden medical images in the steganographic domain, HideMIA makes it difficult for attackers to notice the presence of medical images. Specifically, we propose the Mixture-of-Difference-Convolutions (MoDC) and Asymmetric Wavelet Attention (AsyWA) to enable HideMIA to conduct fine-grained analysis on each wavelet sub-band within the steganographic domain, mining features that are specific to medical images. Moreover, to reduce resource consumption on client devices, we design function-aligned knowledge distillation to obtain a lightweight hiding network, namely LightIH. Extensive experiments on six medical datasets demonstrate that our HideMIA achieves superior MIA performance and protective imperceptibility on medical image segmentation and classification.

Primary Subject Area: [Content] Vision and Language

Secondary Subject Area: [Experience] Multimedia Applications

Relevance To Conference:

HideMIA contributes to multimedia/multimodal processing by introducing a novel privacy-enhancing method for sensitive applications based on deep image hiding.
Firstly, our idea of conducting media analysis directly in the steganographic domain can benefit the broader community as it represents a novel approach to privacy protection, which can certainly be adapted to other sensitive multimedia/multimodal tasks.
Secondly, the challenges identified in our image analysis tasks are likely to exist in other media forms as well, and the solutions we have developed could inspire the community to address these challenges.
Finally, the image hiding technique within HideMIA is a special form of image fusion, categorizing it within the field of multimedia/multimodal processing is reasonable.

Submission Number: 233