The field of medical imaging AI is currently undergoing rapid transformations, with methodical research increasingly translated into clinical practice. Despite these successes, research suffers from knowledge silos, hindering collaboration and progress: Existing knowledge is scattered across publications and many details remain unpublished, while privacy regulations restrict data sharing. In the spirit of democratizing of AI, we propose a framework for secure knowledge transfer in the field of medical image analysis. The key to our approach is dataset "fingerprints", structured representations of feature distributions, that enable quantification of task similarity. We tested our approach across 71 distinct tasks and 12 medical imaging modalities by transferring neural architectures, pretraining, augmentation policies, and multi-task learning. According to comprehensive analyses, our method outperforms traditional methods for identifying relevant knowledge and facilitates collaborative model training. Our framework fosters the democratization of AI in medical imaging and could become a valuable tool for promoting faster scientific advancement.

翻译：医疗影像AI领域当前正经历快速变革，系统性研究日益转化为临床实践。尽管取得这些成功，研究仍受知识孤岛制约，阻碍了协作与进展：现有知识分散于各类出版物中，许多细节仍未公开，而隐私法规限制了数据共享。本着AI民主化的精神，我们提出了一个医疗图像分析领域的安全知识迁移框架。该方法的核心在于数据集"指纹"——特征分布的结构化表征，能够实现任务相似性的量化。我们通过迁移神经网络架构、预训练策略、数据增强方案和多任务学习，在71个独立任务和12种医疗影像模态上验证了本方法。综合分析表明，该方法在识别相关知识方面优于传统方法，并能促进协作式模型训练。本框架推动了医疗影像AI的民主化进程，有望成为加速科学进步的重要工具。