Improving Coronavirus (COVID-19) Diagnosis using Deep Transfer Learning

Abstract

Background: Coronavirus disease is an infectious disease caused by a new virus. Exponential growth is not only threatening lives, but also impacting businesses and disrupting travel around the world. Aim: The aim of this work is to develop an efficient diagnosis of COVID-19 disease by differentiating it from viral pneumonia, bacterial pneumonia and healthy cases using deep learning techniques. Method: In this work, we have used pre-trained knowledge to improve the diagnostic performance using transfer learning techniques and compared the performance different CNN architectures. Results: Evaluation results using K-fold (10) showed that we have achieved state of the art performance with overall accuracy of 98.75% on the perspective of CT and X-ray cases as a whole. Conclusion: Quantitative evaluation showed high accuracy for automatic diagnosis of COVID-19. Pre-trained deep learning models develop in this study could be used early screening of coronavirus, however it calls for extensive need to CT or X-rays dataset to develop a reliable application.

OptCoNet: an optimized convolutional neural network for an automatic diagnosis of COVID-19

“…The advantage of the proposed method is that no settings are required while training the images. Consequently, no tuning is needed for different databases, as opposed to other model-based methodologies, as in [8,14]. In this way, the proposed approach can successfully deal with any concealed databases with no particular parameter tuning.…”

“…In this way, the proposed approach can successfully deal with any concealed databases with no particular parameter tuning. The specialists in [8,14,32] accomplished a marginally poorer accuracy with their DLs compared to the proposed OptCoNet because of the absence of primary data in the images. Likewise, the proposed method is compared with recently published works in terms of the accuracy and F1-score.…”

“…Multilevel thresholding and a support vector machine (SVM) were introduced to achieve high accuracy with images of the infected lungs of COVID-19 patients. Rehman et al [14] utilized pretrained information to improve demonstrative execution using TL methods and performed a comparative analysis with different CNN structures.…”

Applications of artificial intelligence in battling against covid-19: A literature review

“…To get an idea about the covid-19 detection transparency, the concept of Gradient Class Attention Map is used to detect the regions where the model paid more attention during the classification. A pre-trained transfer learning technique is used in [170] and compared with different CNN architectures.…”

A study of neural networks point source extraction on simulated Fermi/LATtelescope images

“…Since Krizhevsky et al (2012), neural networks have been invading different areas of research from medical image analysis by Razzak et al (2017) to jet structure analysis in particle physics by Marzani et al (2019). Concerning approximation of the real data by our artificial dataset, we assume that our neural networks do not require for operation any additional information.…”