def _apply_mask(self, output_dir):
try:
transforms.Compose([
LoadNifti(),
ApplyMask(os.path.join(output_dir, "labels.nii.gz"))
])(os.path.join(output_dir, "brainmask.nii.gz"))
except Exception as e:
self.LOGGER.warning(e)
def apply_mask(file, mask):
transform_ = transforms.Compose([ApplyMask(mask)])
return transform_(file)
generated_mrbrains = transform(
"/mnt/md0/Research/DualUNet/Reconstructed_Normalized_MRBrainS_Image_80.nii.gz"
)
generated_abide = transform(
"/mnt/md0/Research/DualUNet/Reconstructed_Normalized_ABIDE_Image_80.nii.gz"
)
segmentation_iseg = transform(
"/mnt/md0/Research/DualUNet/Reconstructed_Segmented_iSEG_Image_80.nii.gz"
)
segmentation_mrbrains = transform(
"/mnt/md0/Research/DualUNet/Reconstructed_Segmented_MRBrainS_Image_80.nii.gz"
)
segmentation_abide = transform(
"/mnt/md0/Research/DualUNet/Reconstructed_Segmented_ABIDE_Image_80.nii.gz"
)
generated_iseg = torch.tensor(ApplyMask(segmentation_iseg)(generated_iseg))
generated_mrbrains = torch.tensor(
ApplyMask(segmentation_mrbrains)(generated_mrbrains))
generated_abide = torch.tensor(
ApplyMask(segmentation_abide)(generated_abide))
train_samples = torch.cat([iseg_inputs, mrbrains_inputs, abide_inputs],
dim=0).view(90, c * d * h * w)
generated_samples = torch.cat(
[generated_iseg, generated_mrbrains, generated_abide],
dim=0).view(3, c * d * h * w)
p_inputs = compute_probs(train_samples)
p_generated = compute_probs(generated_samples)
js_divergence_inputs = compute_js_divergence(p_inputs).item()