COVID-19 Recognition Using Self-Supervised Learning Approach in Three New Computed Tomography Databases

Amgad Muneer, Rao Faizan Ali, Suliman Mohamed Fati, Sheraz Naseer


Pretraining has been a widely used approach in computer vision and natural language processing, as it typically results in significant performance improvements. The most commonly used pretraining method is transfer learning (TL), which uses labeled data to develop a suitable representation network. Recently, a novel technique for pretraining, self-supervised learning (SSL), has shown promise in various applications. Computed Tomography (CT) is a convenient way to diagnose COVID-19 patients during the COVID-19 pandemic phase. However, openly accessible COVID-19 datasets are highly problematic to gain, which hampers the investigation, and the growth of AI-powered diagnosis approaches of COVID-19 depending on CTs. By using an open-source dataset COVID-CT solve this problem, comprising of 349 COVID-19 CT images. This paper uses three CT datasets to introduce a diagnostic approach based on contrastive self-supervised learning (CSSL) and TL. The effectiveness of CSSL demonstrates an improvement in learning representations. Using multi-task learning and leveraging CSSL pretraining by incorporating lung masks and lesion masks, the authors achieved an accuracy of 0.89 with an F1 of 0.90 and AUC of 0.98.

Keywords: transfer learning, self-supervised learning, COVID-19, computed tomography, lung mask, lesion mask.

Full Text:



HUANG L. Serial Quantitative Chest CT Assessment of COVID-19: A Deep Learning Approach. Radiology: Cardiothoracic Imaging, 2020, 2(2): e200075. doi: 10.1148/ryct.2020200075.

LI L. Using Artificial Intelligence to Detect COVID-19 and Community-acquired Pneumonia Based on Pulmonary CT Evaluation of the Diagnostic Accuracy. Radiology, 2020, 296(2): E65–E71. doi: 10.1148/radiol.2020200905.

XU X. Deep Learning System to Screen Coronavirus Disease. Pneumonia, 2019, 29.

ZHENG C. Deep Learning-based Detection for COVID-19 from Chest CT using Weak Label. Infectious Diseases (except HIV/AIDS), 2020. doi: 10.1101/2020.03.12.20027185.

GOZES O. Rapid ai development cycle for the coronavirus (COVID-19) pandemic: Initial results for automated detection and patient monitoring using deep learning ct image analysis.

YING S. Deep learning Enables Accurate Diagnosis of Novel Coronavirus (COVID-19) with CT images. Radiology and Imaging, 2020. doi: 10.1101/2020.02.23.20026930.

SHI F. Large-Scale Screening of COVID-19 from Community-Acquired Pneumonia using Infection Size-Aware Classification. No date, 8.

COHEN JP, MORRISON P, DAO L. COVID-19 Image Data Collection. arXiv:2003.11597 [cs, eess, q-bio], 2020.

CHOWDHURY MEH. Can AI Help in Screening Viral and COVID-19 Pneumonia? IEEE Access, 2020, 8: 132665–132676. doi: 10.1109/ACCESS.2020.3010287.

HUANG G, LIU Z, VAN DER MAATEN L, WEINBERGER KQ. Densely Connected Convolutional Networks. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, 2261–2269. doi: 10.1109/CVPR.2017.243.

HE K, ZHANG X, REN S, SUN J. Deep Residual Learning for Image Recognition. arXiv:1512.03385 [cs], 2015.

DENG J, DONG W, SOCHER R, LI L-J, LI K, FEI-FEI L. ImageNet: A Large-Scale Hierarchical Image Database. No date, 8.

KINGMA DP, BA J. Adam: A Method for Stochastic Optimization. arXiv:1412.6980 [cs], 2017.

NASEER S, ALI RF, MUNEER A, FATI SM. IAmideV-deep: Valine amidation site prediction in proteins using deep learning and pseudo amino acid compositions. Symmetry, 2021, 13(4): 560.

MUNEER A, FATI SM, FUDDAH S. Smart health monitoring system using IoT based smart fitness mirror. Telkomnika, 2020, 18(1): 317-331.

NASEER S, ALI RF, FATI SM, MUNEER A. iNitroY-Deep: Computational Identification of Nitrotyrosine Sites to Supplement Carcinogenesis Studies Using Deep Learning. IEEE Access, 2021, 9: 73624-73640.

FATI SM, MUNEER A, AKBAR NA, TAIB SM. A Continuous Cuffless Blood Pressure Estimation Using Tree-Based Pipeline Optimization Tool. Symmetry, 2021, 13(4): 686.

LAWTON S, VIRIRI S. Detection of COVID-19 from CT Lung Scans Using Transfer Learning. Computational Intelligence and Neuroscience, 2021.

JUN M., CHENG G., YIXIN W., XINGLE A., JIANTAO G., ZIQI Y., MINQING Z., XIN L., XUEYUAN D., SHUCHENG C., HAO W., SEN M., XIAOYU Y., ZIWEI N., CHEN L., LU T., YUNTAO Z., QIONGJIE Z., GUOQIANG D., and JIAN H. COVID-19 CT Lung and Infection Segmentation Dataset (Verson 1.0) [Data set]. Zenodo, 2020.

MEDICAL SEGMENTATION. COVID-19 CT segmentation dataset, 2020.


  • There are currently no refbacks.