Read MHD and raw images, slice, normalize and save them

Statement

Make suggestions normally , I accept . I advise you not to look at it , If you don't like it, don't look , Cheap or not .

Preface

raw What is stored is a 3D Of CT image ,mhd The file stores the corresponding raw File information

My dataset train/low dose The contents of the folder are as follows ,mhd and raw Under the same folder ：

Read mhd、raw And slice

Use SimpleITK library mhd and raw File reading

import os
from tqdm import tqdm

import SimpleITK as sitk
import matplotlib.pyplot as plt

def slice(ori_path: str, pro_path: str):
    for path in os.listdir(ori_path):
        if path.find('mhd') >= 0:
            data_mhd = sitk.ReadImage(os.path.join(ori_path, path))

            spacing = data_mhd.GetSpacing()
            scan = sitk.GetArrayFromImage(data_mhd)
            for i in tqdm(range(len(scan))):
                plt.figure()
                plt.imshow(scan[i], cmap=plt.cm.bone)
                plt.show()
                break  #  Here is just a demonstration of the effect after slicing , So use the break, Normal traversal will be deleted directly break Just go 

call ：

low_dose_path = 'train/low dose'  #  Original dataset path 
low_dose_slice_path = 'train_slice/low_dose_slice'  #  Save the path 

slice(low_dose_path, low_dose_slice_path)

Picture demonstration example ：

---

section 、 normalization 、 preservation

def slice(ori_path: str, pro_path: str):
    id = 0
    for path in os.listdir(ori_path):
        if path.find('mhd') >= 0:
            id += 1
            save_content = os.path.join(pro_path, str(id))
            if os.path.exists(save_content):
                shutil.rmtree(save_content)
            os.makedirs(save_content)

            data_mhd = sitk.ReadImage(os.path.join(ori_path, path))

            spacing = data_mhd.GetSpacing()
            scan = sitk.GetArrayFromImage(data_mhd)
            for i in tqdm(range(len(scan))):
                img = cv2.normalize(scan[i], None, 0, 255, cv2.NORM_MINMAX, cv2.CV_8U)
                img = Image.fromarray(img)
                save_path = os.path.join(save_content, f'{
      id}_{
      i}.png')
                img.save(save_path)

explain ：
   My dataset 3D The pixel value of the image is -3000 To 3000 many , And our normal images are 0-255 Of , So the best way is to use cv2.normalize Normalize to 0-255 Between , And then save .
   If you write a normalized code by yourself , Or use from sklearn.preprocessing import minmax_scale This minmax_scale function , There will be some strange problems , I guess it may be that the loss of normalization is very large , So there will be some problems .

Call as above ：

low_dose_path = 'train/low dose'  #  Original dataset path 
low_dose_slice_path = 'train_slice/low_dose_slice'  #  Save the path 
slice(low_dose_path, low_dose_slice_path)

Preservation effect ：