ABSTRACT
Background: Coronavirus disease (COVID-19) is an infectious dis- ease 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 of 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 developed in this study could be used for early screening of coronavirus;however, it calls for extensive need to CT or X-rays dataset to develop a reliable application. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
ABSTRACT
The flow of novel coronavirus (COVID-19) has affected almost every aspect of human life around the globe. Being the emerging ground and early sufferer of the virus, Wuhan city-data remains a case of multifold significance. Further, it is of notable importance to explore the impact of unique and unprecedented public health response of Chinese authorities—the extreme lockdown of the city. In this research, we investigate the statistical nature of the viral transmission concerning social distancing, extreme quarantine, and robust lockdown interventions. We observed highly convincing and statistically significant evidences in favor of quarantine and social distancing approaches. These findings might help countries, now facing, or likely to face the wave of the virus. We analyzed Wuhan-based data of “number of deaths” and “confirmed cases,” extracted from China CDC weekly database, dated from February 13, 2020, to March 24, 2020. To estimate the underlying group structure, the assembled data is further sub-divided into three blocks, each consists of two weeks. Thus, the complete data set is studied in three phases, such as, phase 1 (Ph 1) = February 13, 2020, to February 26, 2020;phase 2 (Ph 2) = February 27, 2020 to March 11, 2020;and phase 3 (Ph 3) = March 12, 2020 to March 24, 2020. We observed the overall median proportion of deaths in those six weeks remained 0.0127. This estimate is highly influenced by Ph1, when the early flaws of weak health response were still prevalent. Over the time, we witnessed a median decline of 92.12% in the death proportions. Moreover, a non-parametric version of the variability analysis of death data, estimated that the average rank of reported proportions in Ph 3 remained 7, which was 20.5 in Ph 2, and stayed 34.5 in the first phase. Similar patterns were observed, when studying the confirmed cases data. We estimated the overall median of the proportion of confirmed cases in Wuhan as 0.0041, which again, is highly inclined towards Ph 1 and Ph 2. We also witnessed minimum average rank proportions for Ph 3, such as 7, which was noticeably lower than Ph 2, 21.71, and Ph 1, 32.29. Moreover, the varying degree of clustering indicates that the effectiveness of quarantine based policies is time-dependent. In general, the decline in coronavirus transmission in Wuhan significantly coincides with the lockdown.