仅掩码不足：面向医疗AI的多模态去标识化生成式修复
Masking Is Not Enough: Generative Restoration for Multimodal De-Identification in Medical AI
- Kennesaw State University（肯尼索州立大学）
- Miami University（迈阿密大学）
- Georgia State University（佐治亚州立大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
针对医疗图像-文本数据的PHI泄漏问题，提出端到端多模态净化框架ClinX，结合图像侧生成式修复与文本侧渐进式去标识化，在MedVQA中验证其优于仅OCR掩码的性能。
AI中文摘要：
医疗图像-文本数据可通过可见图像内容及伴随文本暴露受保护健康信息（PHI），为隐私保护型医疗AI系统带来障碍，该风险在多模态系统中尤为突出，图像、问题、报告及临床语境可能进入训练、评估或推理流程。现有医疗视觉-语言基准主要关注任务效用，而去标识化方法常与下游推理分开评估。我们提出ClinX，一种面向医疗图像-文本数据的端到端多模态PHI净化框架。ClinX通过光学字符识别（OCR）检测可见标识符，构建二值PHI掩码，并应用ClinX-PRISM，一种带隐私导向后处理的无跳跃生成式修复模块，用于抑制嵌入的标识符。同时，文本侧PHI通过渐进式去标识化层级减少：正则表达式掩码、语境感知掩码及基于重写的净化。我们在医疗视觉问答（MedVQA）中评估ClinX，联合测量图像侧、文本侧及组合去标识化设置下的PHI泄漏与下游效用。结果表明，仅OCR掩码作为独立方案不足，基于修复的净化能更好保留临床相关视觉语境，同时大幅降低可恢复PHI。
英文摘要：
Medical image-text data can expose protected health information (PHI) through both visible image content as well as accompanying text, creating a barrier to privacy-preserving medical AI systems. This risk is especially prominent in multimodal systems, where images, questions, reports, and clinical context may enter training, evaluation, or inference pipelines. Existing medical vision-language benchmarks primarily emphasize task utility, while de-identification methods are often evaluated separately from downstream reasoning. We introduce ClinX, an end-to-end multimodal PHI sanitization framework for medical image-text data. ClinX detects visible identifiers with optical character recognition (OCR), constructs binary PHI masks, and applies ClinX-PRISM, a no-skip generative restoration module with privacy-oriented post-processing for burned-in identifier suppression. In parallel, text-side PHI is reduced through progressive de-identification levels: regex masking, context-aware masking, and rewrite-based sanitization. We evaluate ClinX in medical visual question answering (MedVQA), jointly measuring PHI leakage and downstream utility across image-side, text-side, and combined de-identification settings. Results show that OCR-only masking is not sufficient as a standalone solution, and restoration-based sanitization better preserves clinically relevant visual context while sharply reducing recoverable PHI.
