[论文解读] A cascade network for Detecting COVID-19 using chest x-rays

Dailin Lv, Wuteng Qi+3 more|arXiv (Cornell University)|May 1, 2020

COVID-19 diagnosis using AI参考文献 13被引用 35

一句话总结

本论文提出 Cascade-SEMEnet，是一个两阶段的级联 SEME-ResNet50 与 SEME-DenseNet169，用于从胸部X光检测肺炎类型并对 COVID-19 进行细化分类，同时提供 Grad-CAM 可视化。

ABSTRACT

The worldwide spread of pneumonia caused by a novel coronavirus poses an unprecedented challenge to the world's medical resources and prevention and control measures. Covid-19 attacks not only the lungs, making it difficult to breathe and life-threatening, but also the heart, kidneys, brain and other vital organs of the body, with possible sequela. At present, the detection of COVID-19 needs to be realized by the reverse transcription-polymerase Chain Reaction (RT-PCR). However, many countries are in the outbreak period of the epidemic, and the medical resources are very limited. They cannot provide sufficient numbers of gene sequence detection, and many patients may not be isolated and treated in time. Given this situation, we researched the analytical and diagnostic capabilities of deep learning on chest radiographs and proposed Cascade-SEMEnet which is cascaded with SEME-ResNet50 and SEME-DenseNet169. The two cascade networks of Cascade - SEMEnet both adopt large input sizes and SE-Structure and use MoEx and histogram equalization to enhance the data. We first used SEME-ResNet50 to screen chest X-ray and diagnosed three classes: normal, bacterial, and viral pneumonia. Then we used SEME-DenseNet169 for fine-grained classification of viral pneumonia and determined if it is caused by COVID-19. To exclude the influence of non-pathological features on the network, we preprocessed the data with U-Net during the training of SEME-DenseNet169. The results showed that our network achieved an accuracy of 85.6\% in determining the type of pneumonia infection and 97.1\% in the fine-grained classification of COVID-19. We used Grad-CAM to visualize the judgment based on the model and help doctors understand the chest radiograph while verifying the effectivene.

研究动机与目标

• 在 RT-PCR 资源有限时，通过利用胸部 X 光影像推动快速 COVID-19 筛查。
• 开发一个两阶段级联网络，先将正常、细菌性肺炎和病毒性肺炎分类，然后对病毒性肺炎进行细化分类以检测 COVID-19。
• 通过全局平均池化、SE 注意力模块、较大输入尺寸、CLAHE、MoEx，以及基于 U-Net 的肺部掩模来提升性能。
• 通过 Grad-CAM 可视化为临床决策提供模型可解释性。

提出的方法

• 构建 Cascade-SEMEnet，其中包含 SEME-ResNet50 和 SEME-DenseNet169。
• 使用更大的输入尺寸并结合全局平均池化以保留病灶细节并增大感受野。
• 在 ResNet50 与 DenseNet169 分支中引入 Squeeze-and-Excitation（SE）注意力。
• 应用 CLAHE 以增强对比度，MoEx 数据增强以融合来自不同病理类别的特征。
• 使用 U-Net 分割肺部区域，在训练时移除非病理特征。
• 使用 Grad-CAM 随于胸部X光图像来可视化分类依据。

实验结果

研究问题

• RQ1 一个两阶段级联网络是否能够从胸部 X 光中准确区分正常、细菌性肺炎和病毒性肺炎？
• RQ2 细分的病毒性肺炎阶段是否能在其他病毒性肺炎中识别 COVID-19？
• RQ3 数据增强（MoEx、CLAHE）和注意力机制是否提升胸部 X 光数据的肺炎分类性能？
• RQ4 通过 U-Net 的肺部区域掩模是否通过去除非病理特征来提升诊断准确性？

主要发现

模型	准确率	F1-score
VGG19	69.69%	0.70
ResNet50	72.81%	0.73
DenseNet169	77.50%	0.78
VGG19-GAP	77.81%	0.78
ResNet50-GAP	74.06%	0.74
DenseNet169-GAP	80.94%	0.81
SE-ResNet50	81.59%	0.82
SE-DenseNet169	81.87%	0.81
SEME-ResNet50	85.62%	0.86
SEME-DenseNet169	80.31%	0.81

• 基线模型（VGG19、ResNet50、DenseNet169）在肺炎类型分类上表现出不同的准确率，其中 DenseNet169 在它们中表现最好。
• 加入全局平均池化（GAP）和 SE 注意力提升了准确性和 ROC 指标在各模型中的表现。
• SEME-ResNet50 在测试配置中达到峰值准确率和 ROC 性能（测试集 85.6%； AUC 0.904）；Grad-CAM 可视化将决策区域定位在肺部区域。
• SEME-DenseNet169 也显示出强劲性能（80.3% 的测试准确率，带 GAP/SE 包；AUC 在图中标注）。
• 与基线相比，SEME 配置在病毒性肺炎和 COVID-19 的分类上提供了显著的准确率和 F1-score 增益；Grad-CAM 确认模型的注意力与肺部区域对齐。