一种用于液体超负荷预测的多模态方法：整合肺部超声与临床数据
A Multimodal Approach for Fluid Overload Prediction: Integrating Lung Ultrasound and Clinical Data
AI总结：
本文提出融合肺部超声图像与临床数据的多模态方法，通过跨域注意力机制和伪样本生成缓解类别不平衡，以分类方式预测透析患者液体超负荷，取得88.31%的最高准确率。
摘要：
在透析患者中管理液体平衡至关重要，因为管理不当可能导致严重并发症。本文提出一种多模态方法，将肺部超声图像的视觉特征与临床数据相结合，以增强对体内多余液体的预测。我们的框架采用独立编码器分别提取各模态的特征，并通过跨域注意力机制将其融合，以捕获互补信息。通过将预测构建为分类任务，模型性能显著优于回归方法。结果表明，多模态模型始终优于单模态模型，尤其是在注意力机制优先关注表格数据时。伪样本生成进一步有助于缓解类别不平衡的分类问题，实现了88.31%的最高准确率。本研究强调了多模态学习在透析患者液体超负荷管理中的有效性，为改善临床结局提供了有价值的见解。
英文摘要：
Managing fluid balance in dialysis patients is crucial, as improper management can lead to severe complications. In this paper, we propose a multimodal approach that integrates visual features from lung ultrasound images with clinical data to enhance the prediction of excess body fluid. Our framework employs independent encoders to extract features for each modality and combines them through a cross-domain attention mechanism to capture complementary information. By framing the prediction as a classification task, the model achieves significantly better performance than regression. The results demonstrate that multimodal models consistently outperform single-modality models, particularly when attention mechanisms prioritize tabular data. Pseudo-sample generation further contributes to mitigating the imbalanced classification problem, achieving the highest accuracy of 88.31%. This study underscores the effectiveness of multimodal learning for fluid overload management in dialysis patients, offering valuable insights for improved clinical outcomes.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
