Optimizing Convolutional Neural Network by Using Genetic Algorithm for COVID-19 Detection in Chest X - Ray Image

Coronavirus Diseases 19 (COVID-19) first appeared in the city of Wuhan, China, at the end of 2019 and spread to several countries at a very fast rate. An early diagnosis of COVID-19 is needed to prevent further spread, reduce the number of deaths. Reverse Transcription Polymerase Chain Reaction (RT-PCR) is one of the methods used to detect COVID-19. However, this method takes a long time so another faster method is required. As the alternative, Chest x-rays image can be used to detect specific symptoms related to COVID-19 by automating the diagnosis. Convolutional Neural Network (CNN) is widely used for disease detection based on computer-aided medical images. In this study, we proposed classification model for the detection of COVID-19 by using CNN method, in which the synaptic weight is optimized by genetic algorithm. The genetic algorithm is known for its popularity that is very good to be used in terms of optimization problem because it offers an optimum solution that retains the model performance and there is no stochastic drop of performance. The result shows that, CNN optimized by genetic algorithm with Adam optimizer gives best performance for accuracy, F1-Score and Area Under Curve (AUC) with value 87.5%, 85.1 % and 86.8% respectively.