Preprocessing MRI #15
Description
"HELLO! Thank you very much for your great work. I have a question about preprocessing for the MRI image code. You use 2 adjacent slices:
" img_arr = np.concatenate((img_arr[:, :, 0:1], img_arr[:, :, 0:1], img_arr, img_arr[:, :, -1:], img_arr[:, :, -1:]), axis=-1)
"
but what is the reason for this number? Why is that number necessary? In the case where I simply input a single slice without adjacents, I wouldn't have to use the adapter, right?
For example, I am performing myocardial segmentation and I have good results (DICE=0.87), but I don't know if having more adjacent slices would make my result better or worse."
Greetings!"