最近处理医院影像数据时,遇到的需求:把预处理后的MRI影像的每个序列的所有dicom文件保存成nii文件,并且要保留原始的dicom信息
预处理后的MRI影像序列
image.png
每个序列下有多张dicom文件
image.png
Python代码:把每个序列的所有dicom文件保存成nii文件,并保留原始的tag信息
导入需要的包
#edited by NickYu 2020.10.23
import numpy as np
import SimpleITK as sitk
import pydicom
import h5py
import os
import nibabel as nib
import dicom2nifti
定义函数
def dicom2Nii(folderPath,savefolder):
'''
dicom序列转成3维的nii文件，并保留原始的dicom元数据信息
'''
count_study = 0
for every_study in os.listdir(folderPath):#遍历所有的病历号
count_study +=1
tmp_MR_path = os.path.join(folderPath,every_study,'MR')#DWI ,T2等
_save_path =os.path.join(savefolder,every_study,'MR2Nii')
for every_MRI in os.listdir(tmp_MR_path):#每个病历号下面可能有多次MRI
tmp_MRI_path = os.path.join(tmp_MR_path,every_MRI)
tmp_save_path = os.path.join(_save_path,every_MRI)
if not os.path.exists(tmp_save_path):
os.makedirs(tmp_save_path)
all_Series_path = os.listdir(tmp_MRI_path)
#print(all_Series_path) #['DCE00001', 'DCE00002', 'DCE00003', 'DCE00004', 'DCE00005', 'DCE00006']
for every_DCE_series in all_Series_path:
#print(every_DCE_series) # DCE00001
all_dicoms = os.path.join(tmp_MRI_path,every_DCE_series)
#print(all_dicoms)
# path_read:读取dicom的文件路径 path_save:保存nii的文件路径
def dcm2nii(path_read, path_save):# from CSDN;function: transfer dcm_series into nii file
# GetGDCMSeriesIDs读取序列号相同的dcm文件
series_id = sitk.ImageSeriesReader.GetGDCMSeriesIDs(path_read)
# GetGDCMSeriesFileNames读取序列号相同dcm文件的路径，series[0]代表第一个序列号对应的文件
series_file_names = sitk.ImageSeriesReader.GetGDCMSeriesFileNames(path_read, series_id[0])
#print(len(series_file_names)) #11
series_reader = sitk.ImageSeriesReader()
series_reader.SetFileNames(series_file_names)
image3d = series_reader.Execute()
sitk.WriteImage(image3d, path_save)
path_save = tmp_save_path + "\\"+every_DCE_series+".nii"
if os.path.exists(path_save):
continue
dcm2nii(all_dicoms, path_save) #调用函数执行
定义待读取文件夹路径和nii文件存储路径
if __name__ =="__main__":
folderPath =r"G:\cut_SkinData_Copy\Cut_Skin"
savefolder = r"G:\2020.10.cut_Skin_DATA(nii)"
dicom2Nii(folderPath,savefolder)
说明：本文为原创文章，转载或引用请注明网址；欢迎一起学习交流，有不正确之处欢迎批评指正2210432548@qq.com
