Review Article

AIoT Used for COVID-19 Pandemic Prevention and Control

Table 3

Methods applied to intelligent image diagnosis of COVID-19 and the results of each method.

AuthorsDatabase usedObtained resultsHighlights

Tang et al. [23]247 COVID-19 patients and 152 other pneumonia patientsIn the algorithm model, the average diagnosis time per person has been reduced to 0.4 s.It has high application value.
Jiang and Xu [24]CT images of patients diagnosed with COVID-19 in Zhongnan hospitalThe sensitivity of the intelligence-assisted diagnosis model is 96%.Comprehensive diagnosis accuracy is high.
Umri et al. [25]GitHub and Kaggle websiteThe accuracy is 98%.Compared with VGG-16, the effect of CNN is better and significant.
Gomes et al. [26]Kaggle websiteThe average accuracy is 89.78%; the average sensitivity is 89.79%.Computing costs are lower than those using deep learning techniques.
Narin [27]Kaggle websiteThe highest sensitivity value is 96.35%.It is beneficial to reduce the doctors’ misdiagnosis rate.
Singh and Singh [28]6,500 chest X-raysThe overall accuracy is 95.83%.It is used to diagnose COVID-19 from chest X-ray images.
Sivaramakrishnan et al. [29]CXR images of children aged 1 to 5 years collected at Guangzhou Medical CenterThe highest accuracy is 99.01%.Weighted average performance significantly improves performance.
Hernandez et al. [30]https://www.sirm.org/category/senza-categoria/COVID-19/The accuracy rate is about 90%.It provides a completely new way of thinking.
Wang et al. [31]Chest CT scans of 251 patients with corresponding voxel-grade lobesThe proposed method has an accuracy of 93.3%.It detects the most accurate location of the lesion area.