Figure 2
From: Classification and visual explanation for COVID-19 pneumonia from CT images using triple learning
Overview of network structure. The proposed method is based on the U-Net architecture. The encoder consists of ResNet18, the output has high dimensional features, and the decoder outputs a segmentation map. The features extracted from ResNet18 are fed into two fully convolution networks (FCNs), and we obtain two types of vectors for classification and contrastive learning. The attention module also teaches the information of infection regions for two FCNs. A ground truth of a semantic segmentation includes three categories. A black region is a background category, and blue and red regions are normal and infection regions.
