Department of CSE, National Institute of Technology, Silchar, India.
J Healthc Eng. 2020 Aug 11;2020:8843664. doi: 10.1155/2020/8843664. eCollection 2020.
Coronavirus Disease (COVID19) is a fast-spreading infectious disease that is currently causing a healthcare crisis around the world. Due to the current limitations of the reverse transcription-polymerase chain reaction (RT-PCR) based tests for detecting COVID19, recently radiology imaging based ideas have been proposed by various works. In this work, various Deep CNN based approaches are explored for detecting the presence of COVID19 from chest CT images. A decision fusion based approach is also proposed, which combines predictions from multiple individual models, to produce a final prediction. Experimental results show that the proposed decision fusion based approach is able to achieve above 86% results across all the performance metrics under consideration, with average AUROC and F1-Score being 0.883 and 0.867, respectively. The experimental observations suggest the potential applicability of such Deep CNN based approach in real diagnostic scenarios, which could be of very high utility in terms of achieving fast testing for COVID19.
冠状病毒病（COVID19）是一种快速传播的传染病，目前正在全球范围内引发医疗危机。由于目前基于逆转录-聚合酶链反应（RT-PCR）的检测 COVID19 的测试存在局限性，最近各种研究工作提出了基于放射影像学的想法。在这项工作中，探索了各种基于深度学习卷积神经网络（Deep CNN）的方法，从胸部 CT 图像中检测 COVID19 的存在。还提出了一种基于决策融合的方法，该方法结合了多个单独模型的预测结果，以生成最终预测。实验结果表明，所提出的基于决策融合的方法能够在考虑的所有性能指标上达到 86%以上的结果，平均 AUROC 和 F1-Score 分别为 0.883 和 0.867。实验观察结果表明，这种基于深度学习卷积神经网络的方法在实际诊断场景中具有潜在的适用性，这对于 COVID19 的快速检测可能具有非常高的实用价值。
