Research on COVID-19 X-ray image recognition algorithm

Abstract

With the Covid-19 pandemic, early identification is considered an important measure to fight against this epidemic. Currently, the risk of secondary infection is high, so it is needed a method that can rapidly identify infected individuals. In this research, two algorithms are used, VGG16 and ResNet50, to classify the lung X-ray images from Kaggle and compare the performance of the two algorithms that ResNet50 has better accuracy than VGG16 on the Chest Radiography Database. Next, the CLAHE image enhancement algorithm is introduced in this research to pre-process the images in the dataset. Finally, the two algorithms, VGG16 and Resnet50, are retrained on the processed dataset and compared with their results. And the results show that the two algorithms are improved but ResNet50 still has better accuracy than VGG16. Thus, the CLAHE algorithm is good in this task, which not only significantly enhances the image quality, but also improves the performance on classification of the neural network model.