About attention_map

[jjjharden]

Hi, thank you for your wonderful work. I am reproducing your DSMIL and I noticed that in the attention_map script
color_map = np.zeros((np.amax(pos_arr, 0)[0]+1, np.amax(pos_arr, 0)[1]+1, 3))
，the shape of color_map is based on the maximum of the patches in the package coordinates, but the packet is obtained after filtering. I was thinking why the shape of color_map is not determined based on the dimension of the WSI?