https://blog.csdn.net/qq_18293213/article/details/72423592
Mainly based on this blog above.
1.
First of all! I don't have a TIF file, but I still want to try whether python3 can install the libtiff package, so I have the courage to explore it. pip install libtiff is directly installed successfully.
from libtiff import *
imgstack = TIFF3D.read_image('train-volume.tif')
for i in range(imgstack.shape[0]):
savepath = '../../deform/tarin/' + str(i) + '.tif'
img = TIFF.open(savepath, 'w')
img.write_image(imgstack[i])
However, the algorithm for converting a tif-volume to a single image is not implemented. Because the image of the tif leaflet is crawled from other code.
2.
I still don't understand the meaning of data enhancement introduced here. And the original link given by the author cannot be opened. depressed.
This is problematic.
1) The path problem, about the path of mid..., all have been changed. Because it contains not only the file name, but also the path, it will cause the path to overlap.
2) To create a folder, under the path of deform
3）
When data enhancement is loaded, the memory may collapse due to the large amount of data. Changed the machine and ran.
3.
Package changed
There will be some problems due to the update of tensorflow. Therefore, the calling function is changed.