Comparison of COVID-19 images classification based on different fusion methods

Inhalt:
The COVID-19 has lasted for more than two years, it still affect people's lives. So it is particularly important to quickly and accurately identify covid-19 images. In this paper, we compared the performance of different feature fusion methods base on VGG19 and ResNet50 for classifying COVID-19 images. Firstly, the two basic network models are used to extract the features from multiple levels and then the extracted features are fused using concat and add, respectively. The fused features are extracted through a convolutional pooling module and finally a layer of full connection and SoftMax are used for classification. In addition, COVID-19 images classification are compared using VGG19, ResNet50 models, and two fusion models. Results show that the accuracy, sensitivity and F1 score of the concat fusion model are 96.8%, 100% and 95.1%.