Trying to design a FCNN for kidney segmentation for MR images. Have 40 images (DICOM) and their associated ground truth labels (PNG) mounted to the Google Drive for the Colab notebook that I am building the architecture upon.
For these 40 images and their labels, trying to resize all of them to 256 x 256 using the code below (and it is practically a carbon copy for the label resizing). Receiving a "list index out of range" for both blocks of code but I am aware of the "n-1" rule for the range() call, so it has to be the glob block of code that is listed just below.
from glob import glob
img_path = glob.glob("/content/drive/My Drive/CHOAS_Kidney_Labels/Training_Images/IMG*.dcm")
mask_path = glob.glob("/content/drive/My Drive/CHOAS_Kidney_Labels/Ground_Truth_Training/IMG*.png")
AttributeError: 'function' object has no attribute 'glob'
images=[]
a=[]
for i in range(0,39):
a=nib.load(img_path[i])
a=a.get_data()
print("image (%d) loaded"%(i))
a=resize(a,(a.shape[0],256,256))
a=a[:,:,:]
for j in range(a.shape[0]):
images.append((a[j,:,:]))
IndexError: list index out of range
Because you used from glob import glob, you don't need the glob.glob. You should just call glob without the namespace glob.
Your second error is because img_path is empty. And you're better off doing something like:
for name in img_path:
a=nib.load(name)