Skip to main content
. 2020 Jan 13;2020:7493419. doi: 10.1155/2020/7493419

Figure 4.

Figure 4

Same images as Figure 2 were used to generate these attention maps. Normalized images [0-1] are pseudocoloured based on the level of “attention” the CNN gave to the region in the image modality (low represented by blue, to high represented by red).