Multi-task driven explainable diagnosis of COVID-19 using chest X-ray images
作者：
Highlights：
• Develop COMiT-Net with 4 joint tasks to segment lung and disease regions using CXR.
• COMiT-Net predicts presence of COVID-19 by differentiating them from healthy lungs.
• Extensive comparison with existing deep learning algorithms for each of the 4 tasks.
• Assemble frontal CXR from various sources against labels for 4 different tasks.
• Creating and publicly releasing manual annotations for lung and disease segmentation.
摘要
•Develop COMiT-Net with 4 joint tasks to segment lung and disease regions using CXR.•COMiT-Net predicts presence of COVID-19 by differentiating them from healthy lungs.•Extensive comparison with existing deep learning algorithms for each of the 4 tasks.•Assemble frontal CXR from various sources against labels for 4 different tasks.•Creating and publicly releasing manual annotations for lung and disease segmentation.
论文关键词：X-Ray,COVID-19,Detection,Diagnostics,Deep learning,Explainable artificial intelligence,Multi-task learning
论文评审过程：Received 21 December 2020, Revised 6 July 2021, Accepted 8 August 2021, Available online 21 August 2021, Version of Record 27 September 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2021.108243
