用生成模型合成多样化肺部超声图像，提升罕见病检测效果
Ultrasound Image Synthesis Using Generative AI for Lung Ultrasound Detection
- 通过病变解剖库引导生成，实现真实感肺超声图像合成
- 对常见病检测提升5.6%准确率，罕见病检测提升25%准确率
- 适合医疗数据稀缺场景，尤其改善罕见病检测能力
构建可靠的医疗AI模型需要代表性与多样性的训练数据。在数据不平衡的情况下，模型在常见类别上性能趋于饱和，而在罕见类别上表现较差。为克服这一限制，我们提出DiffUltra，首个能够生成具有广泛病变变异性的真实肺超声（LUS）图像的生成式AI技术。具体而言，我们通过引入的病变-解剖库（Lesion-anatomy Bank）对生成模型进行条件控制，该库从真实患者数据中捕获病变的结构与位置特征，以指导图像合成。实验表明，与仅使用真实患者数据训练的模型相比，DiffUltra在整体一致性（AP）上提升了5.6%；更重要的是，它显著增加了数据多样性并提升了罕见病例的出现频率，使罕见病例（如占数据集10%的大范围肺实变）的检测准确率提升了25%。
原文摘要 · Abstract (English)
Developing reliable healthcare AI models requires training with representative and diverse data. In imbalanced datasets, model performance tends to plateau on the more prevalent classes while remaining low on less common cases. To overcome this limitation, we propose DiffUltra, the first generative AI technique capable of synthesizing realistic Lung Ultrasound (LUS) images with extensive lesion variability. Specifically, we condition the generative AI by the introduced Lesion-anatomy Bank, which captures the lesion's structural and positional properties from real patient data to guide the image synthesis.We demonstrate that DiffUltra improves consolidation detection by 5.6% in AP compared to the models trained solely on real patient data. More importantly, DiffUltra increases data diversity and prevalence of rare cases, leading to a 25% AP improvement in detecting rare instances such as large lung consolidations, which make up only 10% of the dataset.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。