Skip to main content
Visual Computing for Industry, Biomedicine, and Art
Fig. 7 | Visual Computing for Industry, Biomedicine, and Art

Fig. 7

From: Discrimination between leucine-rich glioma-inactivated 1 antibody encephalitis and gamma-aminobutyric acid B receptor antibody encephalitis based on ResNet18

Fig. 7

Visual explanation of model using Grad-CAM. The final heatmap of each image slice was generated from the average of the heatmaps of the image slices across all loops. The hot color overlaps on the MTL and BG. The intensity of the hot color indicates the degree with which the model focused on the corresponding region for the discrimination between LGI1 and GABAB receptor antibody encephalitis. The Z values in the image slices indicated the MNI coordinates of axial slices

Back to article page