Impact of Lung Segmentation on the Diagnosis and Explanation of COVID-19 in Chest X-ray Images

Sensors

George D. C. Cavalcanti¹  Diego Bertolini²  Lucas O. Teixeira³  Yandre M. G. Costa³  Luiz S. Oliveira⁴  Loris Nanni⁵  Rodolfo M. Pereira⁶ 

[1] Centro de Informática, Universidade Federal de Pernambuco, Recife 50740-560, Brazil;Departamento Acadêmico de Ciência da Computação, Universidade Tecnológica Federal do Paraná, Campo Mourão 87301-899, Brazil;Departamento de Informática, Universidade Estadual de Maringá, Maringá 87020-900, Brazil;Departamento de Informática, Universidade Federal do Paraná, Curitiba 81531-980, Brazil;Dipartimento di Ingegneria dell’Informazione, Università degli Studi di Padova, 35122 Padova, Italy;Instituto Federal do Paraná, Pinhais 83330-200, Brazil;

关键词: COVID-19;    chest X-ray;    semantic segmentation;    explainable artificial intelligence;   

DOI  :  10.3390/s21217116

来源: DOAJ

【 摘 要 】

COVID-19 frequently provokes pneumonia, which can be diagnosed using imaging exams. Chest X-ray (CXR) is often useful because it is cheap, fast, widespread, and uses less radiation. Here, we demonstrate the impact of lung segmentation in COVID-19 identification using CXR images and evaluate which contents of the image influenced the most. Semantic segmentation was performed using a U-Net CNN architecture, and the classification using three CNN architectures (VGG, ResNet, and Inception). Explainable Artificial Intelligence techniques were employed to estimate the impact of segmentation. A three-classes database was composed: lung opacity (pneumonia), COVID-19, and normal. We assessed the impact of creating a CXR image database from different sources, and the COVID-19 generalization from one source to another. The segmentation achieved a Jaccard distance of 0.034 and a Dice coefficient of 0.982. The classification using segmented images achieved an F1-Score of 0.88 for the multi-class setup, and 0.83 for COVID-19 identification. In the cross-dataset scenario, we obtained an F1-Score of 0.74 and an area under the ROC curve of 0.9 for COVID-19 identification using segmented images. Experiments support the conclusion that even after segmentation, there is a strong bias introduced by underlying factors from different sources.

【 授权许可】

Unknown