Open waallf opened 4 years ago
Hi @waallf,
I don't have the visualization script at hand, but it's not hard at all, you just need to obtain the output of the last convolution layer from the predicate branch, which is a feature map of shape 1 x C x H x W if you just input one image, then you take average along the channel dimension and the resultant H x W matrix is the visualized heat map that you want.
Ji
Thanks!
Thanks for your code. I want to visual result of the feature map. Can you share the code with me?