文献速递甲状腺与浅表头颈US

基于深度学习的甲状腺超声结节分割:利用图像中标记提高性能

Deep Learning-Based Segmentation of Nodules in Thyroid Ultrasound: Improving Performance by Utilizing Markers Present in the Images.
PubMed 文献 · Ultrasound Med Biol · 2019-11-04

研究利用1278个结节数据,提出并评估两种利用图像中卡尺标记的深度学习甲状腺结节分割方法;仅用近似掩膜训练DSC为85.1%,人工标注网络DSC为90.4%,加入卡尺引导后DSC升至93.1%,且卡尺引导方法在减少人工标注训练例数时可达无引导网络性能。

要点

  • 利用图像中卡尺标记的深度学习分割方法DSC达93.1%,高于仅用近似掩膜(85.1%)和人工标注(90.4%)。
  • 卡尺引导方法在减少人工标注训练例数时,性能可匹配不使用卡尺引导的网络。
  • 训练例数增加时,所有方法的性能均提高。

核心结论

  • 利用图像中卡尺标记引导的深度学习分割方法DSC为93.1%,优于仅用近似掩膜(85.1%)和人工标注(90.4%)。
  • 卡尺引导方法在减少人工标注训练例数时,性能可匹配不使用卡尺引导的网络。

研究设计与样本

  • 研究使用1278个甲状腺结节数据,提出并评估两种利用图像中卡尺标记的深度学习分割方法。
  • 第一种方法基于卡尺生成近似结节掩膜;第二种方法将人工标注与卡尺自动引导相结合。

诊断标准与关键数值

  • 仅用近似结节掩膜训练时DSC为85.1%。
  • 使用人工标注训练的网络DSC为90.4%。
  • 加入卡尺引导后DSC升至93.1%。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

英文原文(摘要)· 点击展开

Computer-aided segmentation of thyroid nodules in ultrasound imaging could assist in their accurate characterization. In this study, using data for 1278 nodules, we proposed and evaluated two methods for deep learning-based segmentation of thyroid nodules that utilize calipers present in the images. The first method used approximate nodule masks generated based on the calipers. The second method combined manual annotations with automatic guidance by the calipers. When only approximate nodule masks were used for training, the achieved Dice similarity coefficient (DSC) was 85.1%. The performance of a network trained using manual annotations was DSC = 90.4%. When the guidance by the calipers was added, the performance increased to DSC = 93.1%. An increase in the number of cases used for training resulted in increased performance for all methods. The proposed method utilizing the guidance by calipers matched the performance of the network that did not use it with a reduced number of manually annotated training cases.

来源:Ultrasound Med Biol,查看原文。版权归原作者所有。

引用

  1. Buda M, Wildman-Tobriner B, Castor K, et al. Deep Learning-Based Segmentation of Nodules in Thyroid Ultrasound: Improving Performance by Utilizing Markers Present in the Images. Ultrasound Med Biol. 2019. PMID: 31699547. doi:10.1016/j.ultrasmedbio.2019.10.003

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。