Scribd
Upload
274 views6 pages

DICOM Image Pre-Processing

This document discusses pre-processing of DICOM images. It imports necessary libraries, defines functions to extract information from DICOM files and display pixel arrays. It then loads a DICOM scan, extracts pixel data in Hounsfield units, saves the images and displays a histogram of Hounsfield values. Sample code shows information extracted from a DICOM file and displays its pixel array.

Uploaded by

Dong Hai Nguyen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
274 views6 pages

DICOM Image Pre-Processing

This document discusses pre-processing of DICOM images. It imports necessary libraries, defines functions to extract information from DICOM files and display pixel arrays. It then loads a DICOM scan, extracts pixel data in Hounsfield units, saves the images and displays a histogram of Hounsfield values. Sample code shows information extracted from a DICOM file and displays its pixel array.

Uploaded by

Dong Hai Nguyen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

DICOM Image Pre-processing

July 8, 2019

[1]: import numpy as np


import pydicom
import os
import matplotlib.pyplot as plt
from glob import glob
from IPython.display import Image
[2]: def show_dcm_info(dataset):
print("Filename.........:", file_path)
print("Storage type.....:", dataset.SOPClassUID)
print()

pat_name = dataset.PatientName
display_name = pat_name.family_name + ", " + pat_name.given_name
print("Patient's name......:", display_name)
print("Patient id..........:", dataset.PatientID)
print("Patient's Age.......:", dataset.PatientAge)
print("Patient's Sex.......:", dataset.PatientSex)
print("Modality............:", dataset.Modality)
print("Body Part Examined..:", dataset.BodyPartExamined)
print("View Position.......:", dataset.ViewPosition)

if 'PixelData' in dataset:
rows = int(dataset.Rows)
cols = int(dataset.Columns)
print("Image size.......: {rows:d} x {cols:d}, {size:d} bytes".format(
rows=rows, cols=cols, size=len(dataset.PixelData)))
if 'PixelSpacing' in dataset:
print("Pixel spacing....:", dataset.PixelSpacing)
[3]: def plot_pixel_array(dataset, figsize=(10,10)):
plt.figure(figsize=figsize)
plt.imshow(dataset.pixel_array, cmap=plt.cm.bone)
# python -m pip install pillow
# pip install glob3
plt.show()

1
[4]: file_path=""
data_path = "C:
,→\\Users\\dongh\\Documents\\siim-acr-pneumothorax-segmentation\\sample_images"

output_path = working_path = "C:\\Users\\dongh\\"


g = glob(data_path + '/*.dcm')

# Print out the first 5 file names to verify we're in the right folder.
print ("Total of %d DICOM images.\nFirst 5 filenames:" % len(g))
print ('\n'.join(g[:5]))

Total of 1 DICOM images.


First 5 filenames:
C:\Users\dongh\Documents\siim-acr-pneumothorax-segmentation\sample_images\1.2.27
6.0.7230010.3.1.4.8323329.12743.1517875241.599591.dcm

[5]: #
# Loop over the image files and store everything into a list.
#

def load_scan(path):
slices = [pydicom.read_file(path + '/' + s,force=True) for s in os.
,→listdir(path)]

slices.sort(key = lambda x: int(x.InstanceNumber))


print (len(slices))

try:
slice_thickness = np.abs(slices[0].PixelSpacing[0] - slices[0].
,→PixelSpacing[1])

except:
slice_thickness = np.abs(slices[0].PixelSpacing[1] - slices[0].
,→PixelSpacing[2])

for s in slices:
s.SliceThickness = slice_thickness
return slices

def get_pixels_hu(scans):
image = np.stack([s.pixel_array for s in scans])
# Convert to int16 (from sometimes int16),
# should be possible as values should always be low enough (<32k)
image = image.astype(np.int16)

# Set outside-of-scan pixels to 1


# The intercept is usually -1024, so air is approximately 0
image[image == -2000] = 0

2
# Convert to Hounsfield units (HU)
if ('RescaleIntercept' in scans) and ('RescaleSlope' in scans):
intercept = scans[0].RescaleIntercept
slope = scans[0].RescaleSlope

if slope != 1:
image = slope * image.astype(np.float64)
image = image.astype(np.int16)

image += np.int16(intercept)

return np.array(image, dtype=np.int16)

id=0
patient = load_scan(data_path)
imgs = get_pixels_hu(patient)

[6]: np.save(output_path + "fullimages_%d.npy" % (id), imgs)


[7]: # Hounsfield-Units-for-human-body.png
Image("img/Hounsfield-Units-for-human-body.png")
[7]:

3
[8]: file_used=output_path+"fullimages_%d.npy" % id
imgs_to_process = np.load(file_used).astype(np.float64)

plt.hist(imgs_to_process.flatten(), bins=50, color='c')


plt.xlabel("Hounsfield Units (HU)")
plt.ylabel("Frequency")
plt.show()

[9]: # Validation
Image("img/dicom_view.png")
# actual image
[9]:

4
[10]: samples_files = glob('siim-acr-pneumothorax-segmentation\sample_images\*.dcm')

for file_path in samples_files :


dataset = pydicom.dcmread(file_path)
show_dcm_info(dataset)
plot_pixel_array(dataset)
break;
# Python rendering image

Filename...: siim-acr-pneumothorax-segmentation\sample_images\1.2.276.0.72
30010.3.1.4.8323329.12743.1517875241.599591.dcm
Storage type...: 1.2.840.10008.5.1.4.1.1.7

Patient's name...: fcb8a3cf-9449-4186-b29c-db5b93c9c75c,


Patient id...: fcb8a3cf-9449-4186-b29c-db5b93c9c75c
Patient's Age...: 28
Patient's Sex...: M
Modality...: CR

5
Body Part Examined..: CHEST
View Position...: PA
Image size...: 1024 x 1024, 112200 bytes
Pixel spacing...: ['0.19431099999999998', '0.19431099999999998']

You might also like