PyRadiomics toolbox in Python

In this notebook, we will load 2 brain datasets with segmentations of brain tumors. The radiomics package will be used to extract a set of features, and the "signatures" will be compared.

For more detailed examples and links to source code, visit http://radiomics.io



In [7]:

    
# Radiomics package
from radiomics import featureextractor

import six, numpy as np

Setting up data

Here we use SimpleITK (referenced as sitk, see http://www.simpleitk.org/ for details) to load two brain images and the corresponding segmentations as label maps.



In [8]:

    
import os
import SimpleITK as sitk

from radiomics import getTestCase

imagepath_1, labelpath_1 = getTestCase('brain1')
imagepath_2, labelpath_2 = getTestCase('brain2')

image_1 = sitk.ReadImage(imagepath_1)
label_1 = sitk.ReadImage(labelpath_1)
image_2 = sitk.ReadImage(imagepath_2)
label_2 = sitk.ReadImage(labelpath_2)

Show the images

Using matplotlib.pyplot (referenced as plt), display the images in grayscale and labels in color.



In [9]:

    
# Display the images
%matplotlib inline
import matplotlib.pyplot as plt

plt.figure(figsize=(20,20))
# First image
plt.subplot(2,2,1)
plt.imshow(sitk.GetArrayFromImage(image_1)[12,:,:], cmap="gray")
plt.title("Brain #1")
plt.subplot(2,2,2)
plt.imshow(sitk.GetArrayFromImage(label_1)[12,:,:])        
plt.title("Segmentation #1")

# Second image
plt.subplot(2,2,3)
plt.imshow(sitk.GetArrayFromImage(image_2)[8,:,:], cmap="gray")        
plt.title("Brain #2")

plt.subplot(2,2,4)
plt.imshow(sitk.GetArrayFromImage(label_2)[8,:,:])        
plt.title("Segmentation #2")

plt.show()

Extract the features

Using the radiomics package, first construct an extractor object from the parameters set in Params.yaml. Then generate features for the two images.



In [10]:

    
import os
# Instantiate the extractor
params = os.path.join(os.getcwd(), '..', 'examples', 'exampleSettings', 'Params.yaml')

extractor = featureextractor.RadiomicsFeatureExtractor(params)
result_1 = extractor.execute(image_1, label_1)
result_2 = extractor.execute(image_2, label_2)

Prepare for plotting

Because we'd like to plot the feature vectors, create numpy arrays for features starting with original_ (excluding meta-features).



In [11]:

    
# Make an array of the values
feature_1 = np.array([])
feature_2 = np.array([])

for key, value in six.iteritems(result_1):
    if key.startswith("original_"):
        feature_1 = np.append ( feature_1, result_1[key])
        feature_2 = np.append ( feature_2, result_2[key])

Plot

Plot the two feature vectors and the difference. Feature values have a wide range of magnitudes and are plotted on a log scale.



In [12]:

    
plt.figure(figsize=(20,20))
plt.subplot(3,1,1)
plt.plot(feature_1)
plt.yscale('log')
plt.title ( "Features from brain #1")
        
plt.subplot(3,1,2)
plt.plot(feature_2)
plt.yscale('log')
plt.title ( "Features from brain #1")

plt.subplot(3,1,3)
plt.plot(feature_1 - feature_2)
plt.yscale('log')
plt.title ( "Difference")
plt.show()