hi, i used your implementation of U-net to segment Nifti sequences that contain stroke lesions, i didn't change anything, and i got the same result as you, but when i try to visualize the predicted masks and the ground truth mask and images, there's no correspondings between them(in the predict.py instead of visualizing the combined_predicted i visualize the y-predicted, and it's completely different of the image x, also instead of visualization x+y which is the groundtruth+original image i visualized only y, and the mask is not the corresponding mask of x)
hi, i used your implementation of U-net to segment Nifti sequences that contain stroke lesions, i didn't change anything, and i got the same result as you, but when i try to visualize the predicted masks and the ground truth mask and images, there's no correspondings between them(in the predict.py instead of visualizing the combined_predicted i visualize the y-predicted, and it's completely different of the image x, also instead of visualization x+y which is the groundtruth+original image i visualized only y, and the mask is not the corresponding mask of x)
what you think