To understand gradient-weighted class activation mapping (Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization, https://arxiv.org/abs/1610.02391), let's quote the source paper itself:
"Grad-CAM uses the gradients of any target concept (say, the logits for 'dog' or even a caption) flowing into the final convolutional layer to produce a coarse localization map highlighting the important regions in the image for predicting the concept."
The following screenshot shows the Grad-CAM algorithm:
Grad-CAM schema; Source: https://arxiv.org/abs/1610.02391
Now, let's look at how it ...
