scripts/DL_save_nifti.py

"""
A simple demo to load 2D 16-bit slices from DeepLesion and save to 3D nifti volumes.
The nifti volumes can be viewed in software such as 3D slicer and ITK-SNAP.
"""

import os
import cv2

import numpy as np
import pandas as pd
import SimpleITK as sitk


dir_in = '../Images_png'
dir_out = '../Images_nifti'
info_fn = '../DL_info.csv'


def slices2nifti(ims, fn_out, spacing):
    """save 2D slices to 3D nifti file considering the spacing"""
    image_itk = sitk.GetImageFromArray(np.stack(ims, axis=0))
    image_itk.SetSpacing(spacing)
    image_itk.SetDirection((1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, -1.0))
    sitk.WriteImage(image_itk, os.path.join(dir_out, fn_out))
    print(fn_out, 'saved')


def load_slices(dir, slice_idxs):
    """load slices from 16-bit png files"""
    slice_idxs = np.array(slice_idxs)
    assert np.all(slice_idxs[1:] - slice_idxs[:-1] == 1)
    ims = []
    for slice_idx in slice_idxs:
        path = os.path.join(dir_in, dir, f'{slice_idx:03d}.png')
        im = cv2.imread(path, cv2.IMREAD_UNCHANGED)  # Read as 16-bit image
        assert im is not None, f'error reading {path}'
        print(f'read {path}')

        # the 16-bit png file has an intensity bias of 32768
        ims.append((im.astype(np.int32) - 32768).astype(np.int16))
    return ims


if __name__ == '__main__':

    # Read spacings and image indices in DeepLesion
    dl_info = pd.read_csv(info_fn)
    idxs = dl_info[['Patient_index', 'Study_index', 'Series_ID']].values
    spacings = dl_info['Spacing_mm_px_'].apply(lambda x: np.array(x.split(", "), dtype=float)).values
    spacings = np.stack(spacings)

    if not os.path.exists(dir_out):
        os.mkdir(dir_out)
    img_dirs = sorted(os.listdir(dir_in))
    for dir1 in img_dirs:
        # find the image info according to the folder's name
        idxs1 = np.array([int(d) for d in dir1.split('_')])
        i1 = np.where(np.all(idxs == idxs1, axis=1))[0]
        spacings1 = spacings[i1[0]]

        fns = os.listdir(os.path.join(dir_in, dir1))
        slices = sorted([int(d[:-4]) for d in fns if d.endswith('.png')])

        # Each folder contains png slices from one series (volume)
        # There may be several sub-volumes in each volume depending on the key slices
        # We group the slices into sub-volumes according to continuity of the slice indices
        groups = []
        for slice_idx in slices:
            if len(groups) != 0 and slice_idx == groups[-1][-1]+1:
                groups[-1].append(slice_idx)
            else:
                groups.append([slice_idx])

        for group in groups:
            # group contains slices indices of a sub-volume
            ims = load_slices(dir1, group)
            fn_out = f'{dir1}_{group[0]:03d}-{group[-1]:03d}.nii.gz'
            slices2nifti(ims, fn_out, spacings1)