基于计算机断层扫描图像的快速鲁棒股骨分割用于患者特异性髋骨折风险筛查
图像与视频处理
2022-07-06 v1 计算机视觉与模式识别
摘要
骨质疏松症是一种常见骨病，会增加骨折风险。基于有限元分析的髋骨折风险筛查方法依赖于分割后的计算机断层扫描（CT）图像；然而，当前的股骨分割方法需要对大型数据集进行手动勾画。在此，我们提出一种深度神经网络，用于从 CT 中对股骨近端进行全自动、准确且快速的分割。在包含 1147 个具有真实分割的股骨近端样本集上的评估表明，我们的方法适用于髋骨折风险筛查，使我们向临床上可行的风险患者髋骨折易感性筛查方案更近一步。
引用
@article{arxiv.2204.09575,
title = {Fast and Robust Femur Segmentation from Computed Tomography Images for Patient-Specific Hip Fracture Risk Screening},
author = {Pall Asgeir Bjornsson and Alexander Baker and Ingmar Fleps and Yves Pauchard and Halldor Palsson and Stephen J. Ferguson and Sigurdur Sigurdsson and Vilmundur Gudnason and Benedikt Helgason and Lotta Maria Ellingsen},
journal= {arXiv preprint arXiv:2204.09575},
year = {2022}
}
备注
This article has been accepted for publication in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, published by Taylor & Francis