[论文解读] COVID-Net: A Tailored Deep Convolutional Neural Network Design for Detection of COVID-19 Cases from Chest X-Ray Images

Linda Wang, Alexander Wong|arXiv (Cornell University)|Mar 22, 2020

COVID-19 diagnosis using AI参考文献 48被引用 267

一句话总结

COVID-Net 是一个开源、定制的 CNN，用于从胸部 X 光图像检测 COVID-19，采用人机协同设计过程构建，并在 COVIDx 上进行评估，其可解释性审计验证了其决策依据。

ABSTRACT

The COVID-19 pandemic continues to have a devastating effect on the health and well-being of the global population. A critical step in the fight against COVID-19 is effective screening of infected patients, with one of the key screening approaches being radiology examination using chest radiography. Motivated by this and inspired by the open source efforts of the research community, in this study we introduce COVID-Net, a deep convolutional neural network design tailored for the detection of COVID-19 cases from chest X-ray (CXR) images that is open source and available to the general public. To the best of the authors' knowledge, COVID-Net is one of the first open source network designs for COVID-19 detection from CXR images at the time of initial release. We also introduce COVIDx, an open access benchmark dataset that we generated comprising of 13,975 CXR images across 13,870 patient patient cases, with the largest number of publicly available COVID-19 positive cases to the best of the authors' knowledge. Furthermore, we investigate how COVID-Net makes predictions using an explainability method in an attempt to not only gain deeper insights into critical factors associated with COVID cases, which can aid clinicians in improved screening, but also audit COVID-Net in a responsible and transparent manner to validate that it is making decisions based on relevant information from the CXR images. By no means a production-ready solution, the hope is that the open access COVID-Net, along with the description on constructing the open source COVIDx dataset, will be leveraged and build upon by both researchers and citizen data scientists alike to accelerate the development of highly accurate yet practical deep learning solutions for detecting COVID-19 cases and accelerate treatment of those who need it the most.

研究动机与目标

• 在资源有限的环境中，使用胸部 X 光影像实现对 COVID-19 的快速、可及筛查的动机。
• 介绍 COVID-Net，一个为 COVID-19 检测的胸 X 光影像定制的 CNN。
• 创建 COVIDx，一个面向 COVID-19 研究的大型开放获取的胸 X 光基准数据集。
• 启用可解释性分析以审计模型决策并提升临床信任。

提出的方法

• 为三元预测（正常、非 COVID-19 感染、COVID-19）开发初始的人为设计残差原型。
• 使用机器驱动的设计探索（生成式综合）以 COVID-19 敏感性 ≥ 80% 和 PPV ≥ 80% 作为约束。
• 使用轻量化的投影–扩张–投影（PEPX）块和有选择的长程连接来构建 COVID-Net。
• 在 ImageNet 上进行预训练并在 COVIDx 上使用 Adam 进行训练，学习率 2e-4，22 轮，批量大小 64，并加上数据增强和批量再平衡。
• 使用 GSInquire 可解释性对模型决策进行审计，以确保依赖于相关肺部区域而非伪影。

实验结果

研究问题

• RQ1 COVID-Net 是否能够在 COVIDx 数据集上实现对 COVID-19 检测的高灵敏度和高 PPV？
• RQ2 来自机器驱动探索的架构设计选择（PEPX 块、选择性连接和架构多样性）是否提高了与标准架构相比的性能和效率？
• RQ3 可解释性（GSInquire）在揭示 COVID-Net 预测背后的真实因素以及验证可信决策方面有多高的揭示力？

主要发现

• COVID-Net 在 COVIDx 上实现了 93.3% 的测试准确率。
• COVID-Net 对 COVID-19 的灵敏度达到 91.0%，对 COVID-19 的 PPV 达到 98.9%。
• COVID-Net 使用大约 1175 万参数和大约 7.50 G MACs，复杂度远低于 VGG-19 和 ResNet-50。
• 在对比中，COVID-Net 的 COVID-19 灵敏度和 PPV 高于 VGG-19 和 ResNet-50。
• GSInquire 可解释性显示 COVID-Net 依赖于肺部区域进行检测，支持可信推理。