COVID19XrayNet: A Two-Step Transfer Learning Model for the COVID-19 Detecting Problem Based on a Limited Number of Chest X-Ray Images

• Published in: Interdisciplinary sciences : computational life sciences Vol. 12; no. 4; pp. 555 - 565
• Main Authors: Zhang, Ruochi, Guo, Zhehao, Sun, Yue, Lu, Qi, Xu, Zijian, Yao, Zhaomin, Duan, Meiyu, Liu, Shuai, Ren, Yanjiao, Huang, Lan, Zhou, Fengfeng
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.12.2020; Springer Nature B.V
• Subjects: Algorithms; Artificial neural networks; Betacoronavirus; Biomedical and Life Sciences; Chest; Clinical Laboratory Techniques - methods; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Appl. in Life Sciences; Coronaviridae; Coronavirus; Coronavirus Infections - complications; Coronavirus Infections - diagnosis; Coronavirus Infections - diagnostic imaging; Coronavirus Infections - virology; Coronaviruses; COVID-19; COVID-19 Testing; Databases, Factual; Datasets; Datasets as Topic; Deep Learning; Diagnostic systems; Health Sciences; Humans; Life Sciences; Lung - diagnostic imaging; Lung diseases; Machine Learning; Mathematical and Computational Physics; Medical imaging; Medicine; Model accuracy; Models, Biological; Neural networks; Neural Networks, Computer; Pandemics; Pneumonia - diagnosis; Pneumonia - diagnostic imaging; Pneumonia - etiology; Pneumonia - virology; Pneumonia, Viral - complications; Pneumonia, Viral - diagnosis; Pneumonia, Viral - diagnostic imaging; Pneumonia, Viral - virology; Radiography - methods; Reference Values; SARS-CoV-2; Severe acute respiratory syndrome coronavirus 2; Short Communication; Statistics for Life Sciences; Theoretical; Theoretical and Computational Chemistry; Tomography, X-Ray Computed - methods; Training; Transfer learning; Viral diseases; X ray imagery; X-Rays; Feature smoothing layer (FSL); ResNet34; Two-step transfer learning; Feature extraction layer (FEL); COVID19XrayNet
• DOI: 10.1007/s12539-020-00393-5
• ISSN: 1913-2751; 1867-1462

The novel coronavirus severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused a major pandemic outbreak recently. Various diagnostic technologies have been under active development. The novel coronavirus disease (COVID-19) may induce pulmonary failures, and chest X-ray imaging becomes one of the major confirmed diagnostic technologies. The very limited number of publicly available samples has rendered the training of the deep neural networks unstable and inaccurate. This study proposed a two-step transfer learning pipeline and a deep residual network framework COVID19XrayNet for the COVID-19 detection problem based on chest X-ray images. COVID19XrayNet firstly tunes the transferred model on a large dataset of chest X-ray images, which is further tuned using a small dataset of annotated chest X-ray images. The final model achieved 0.9108 accuracy. The experimental data also suggested that the model may be improved with more training samples being released. Graphic abstract COVID19XrayNet, a two-step transfer learning framework designed for biomedical images.