Online COVID-19 Computer-aided Diagnostic System with Convolution Neural Network
Conference: CAIBDA 2022 - 2nd International Conference on Artificial Intelligence, Big Data and Algorithms
06/17/2022 - 06/19/2022 at Nanjing, China
Proceedings: CAIBDA 2022
Pages: 6Language: englishTyp: PDF
Authors:
Fu, Qiqi (Faculty of Marine Science and Technology Dalian University of Technology Dalian, China)
Liu, Andi (School of Telecommunication Engineering Xidian University Xi’an, China)
Wu, Zhuolin (School of Engineering Technology Beijing Normal University, Zhuhai, China)
Zhou, Yunpeng (Faculty of Innovation Engineering M.U.S.T Macau, China)
Abstract:
Rapid and accurate identification of positive COVID-19 cases can effectively avoid the widespread of COVID-19, which contributes to fighting the epidemic. Current COVID-19 detection systems use Lung CT shadows for effective detection. However, these systems often exist as offline software, thus placing high demands on the energy and hardware of the institutions using them. This paper establishes an online CT-assisted diagnostic system called COVID-19 detection system that helps doctors judge positive patients faster and more accurately. Firstly, we design a lung CT image recognition model based on CNN. Secondly, we collect data from Github and divide the data into a training set and a test set with a ratio of 4:1 for training and testing. We also use the Relu activation function, Adam algorithm, and data augmentation to optimize the model. Finally, we deploy our model on a web page based on the Flask framework for online use. The COVID-19 detection system is functioning efficiently to detect positive cases of COVID-19. To verify the effectiveness of the proposed system, we test the accuracy of our model on the testing dataset and the testing accuracy is 81.6%. Furthermore, we also test the runtime of the COVID-19 detection system. The COVID-19 detection system can predict the input image in 1.5-3 seconds.