[论文解读] A novel unsupervised covid lung lesion segmentation based on the lung tissue identification
本文提出了一种无监督深度学习框架,通过利用两个独立训练的残差网络——DL-Covid 和 DL-Norm——分别对感染患者和正常患者的肺组织进行识别,实现对 COVID-19 肺部病变在 CT 扫描中的分割。病变图通过两个概率图相减生成,在人工分割基准上达到 0.67 的 Dice 系数和 0.60 的 Jaccard 指数,表现出色,且无需使用标注的病变数据。
This study aimed to evaluate the performance of a novel unsupervised deep learning-based framework for automated infections lesion segmentation from CT images of Covid patients. In the first step, two residual networks were independently trained to identify the lung tissue for normal and Covid patients in a supervised manner. These two models, referred to as DL-Covid and DL-Norm for Covid-19 and normal patients, respectively, generate the voxel-wise probability maps for lung tissue identification. To detect Covid lesions, the CT image of the Covid patient is processed by the DL-Covid and DL-Norm models to obtain two lung probability maps. Since the DL-Norm model is not familiar with Covid infections within the lung, this model would assign lower probabilities to the lesions than the DL-Covid. Hence, the probability maps of the Covid infections could be generated through the subtraction of the two lung probability maps obtained from the DL-Covid and DL-Norm models. Manual lesion segmentation of 50 Covid-19 CT images was used to assess the accuracy of the unsupervised lesion segmentation approach. The Dice coefficients of 0.985 and 0.978 were achieved for the lung segmentation of normal and Covid patients in the external validation dataset, respectively. Quantitative results of infection segmentation by the proposed unsupervised method showed the Dice coefficient and Jaccard index of 0.67 and 0.60, respectively. Quantitative evaluation of the proposed unsupervised approach for Covid-19 infectious lesion segmentation showed relatively satisfactory results. Since this framework does not require any annotated dataset, it could be used to generate very large training samples for the supervised machine learning algorithms dedicated to noisy and/or weakly annotated datasets.
研究动机与目标
- 开发一种无监督深度学习框架,用于自动分割 CT 扫描中的 COVID-19 肺部病变。
- 通过仅依赖肺组织识别,消除对昂贵且耗时的手动病变标注的依赖。
- 通过自监督的概率图相减方法,提升病变分割模型的泛化能力和可扩展性。
- 在 50 例经人工分割的 COVID-19 CT 扫描数据集上验证该方法,以确保其临床相关性。
提出的方法
- 两个独立的残差网络(DL-Covid 和 DL-Norm)以监督方式训练,分别生成 COVID-19 患者和正常患者肺组织的体素级概率图。
- 在正常患者上训练的 DL-Norm 模型,由于对病理模式不熟悉,对感染区域产生较低的概率评分。
- 通过在每个体素上将 DL-Norm 概率图从 DL-Covid 概率图中减去,计算病变概率图。
- 减法过程放大了组织特征的差异,突出显示具有异常强度和纹理的区域,这些区域提示感染。
- 该方法仅依赖预训练模型的输出,推理过程中无需任何病变级别的标注。
- 在包含 50 例 COVID-19 CT 扫描的测试集上进行了外部验证,通过人工分割进行定量评估。
实验结果
研究问题
- RQ1无监督深度学习方法是否能在无需病变级别标注的情况下,实现对 COVID-19 CT 扫描中肺部病变的准确分割?
- RQ2对在正常患者和感染患者上训练的模型的概率图进行相减,对病变检测的有效性如何?
- RQ3所提出的方法是否能在未见病例中实现高 Dice 和 Jaccard 相似度,表现出良好的泛化能力?
- RQ4当将该方法应用于具有不同疾病严重程度的真实临床 CT 数据时,其性能如何?
主要发现
- 所提出的无监督方法在验证数据集上的病变分割中达到 0.67 的 Dice 系数和 0.60 的 Jaccard 指数。
- 肺部分割性能优异,外部验证中正常病例的 Dice 得分为 0.985,COVID-19 病例为 0.978。
- 该方法通过利用在正常患者和感染患者上训练的模型之间的概率图差异,有效识别出病变区域。
- 该框架无需任何标注的病变数据,可实现未来监督模型训练数据的可扩展生成。
- 基于减法的策略成功突出了病理区域,且不依赖于疾病特异性监督。
更好的研究,从现在开始
从阅读论文到最终审阅,大幅缩短您的研究时间。
无需绑定信用卡
本解读由 AI 生成,并经人工编辑审核。