中文

液体超负荷预测的多模态方法:结合肺部超声与临床数据

图像与视频处理 2024-10-04 v2

摘要

管理透析患者的液体平衡至关重要,因为管理不当可能导致严重并发症。我们提出了一种多模态方法,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。我们的框架为每种模态提取特征,并通过跨模态注意力机制进行融合,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处理表格数据时。此外,伪样本生成有助于缓解类别不平衡问题,最终达到了88.31%的最高准确率。这项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宝贵见解。

关键词

引用

@article{arxiv.2409.08790,
  title  = {A Multimodal Approach for Fluid Overload Prediction: Integrating Lung Ultrasound and Clinical Data},
  author = {Tianqi Yang and Nantheera Anantrasirichai and Oktay Karakuş and Marco Allinovi and Alin Achim},
  journal= {arXiv preprint arXiv:2409.08790},
  year   = {2024}
}

备注

In the experiment, for the classification tasks, the network was informed with ground truth during training, significantly improving the performance. This makes the results invalid. Therefore, corrections and more validations are needed to evaluate the performance of the method