Figure 6
From: Classification and visual explanation for COVID-19 pneumonia from CT images using triple learning
Results of visual explanation. (a) Input image, (b) baseline with Grad-CAM23, (c) WSDL with CAM4, (d) attention map using an ABN44, and (e) the attention map achieved by Triple Net (ours). All explanation images are a superposition of input images and heat maps. Red shows the most important location, and blue indicates an unimportant location for classification.
