Skip to main content
. 2020 Apr 10;20(7):2145. doi: 10.3390/s20072145

Figure 12.

Figure 12

(a) the 32 3×3 filters of Conv1 of the network, (b) the input image (cyan circles are marked manually for a better visualization), and (ce), one of the feature activations from the 80th, 86th, and 92th convolutional layers, respectively.