Helper Module for Deep Learning.

pynet data preprocessing¶

Credit: A Grigis

pynet contains a set of tools to efficiently preprocess 3D medical images that is crutial for deep learning applications. It includes reorientation, bias field correction, affine alignement, and intensity normalization.

Load the data¶

We load the Brats dataset and select the first MRI brain image.

import os
import sys
if "CI_MODE" in os.environ:
    sys.exit()

import time
import numpy as np
import random
import nibabel
from pynet.datasets import fetch_toy
from pynet.preprocessing import rescale

datasetdir = "/tmp/toy"
data = fetch_toy(datasetdir=datasetdir)
image = nibabel.load(data.t1w_path)
target = nibabel.load(data.template_path)

Define preprocessing steps¶

We now declare MRI brain preprocessing functions that can be combined with the Processor class.

from pynet.preprocessing import zscore_normalize
from pynet.preprocessing import kde_normalize
from pynet.preprocessing import reorient2std
from pynet.preprocessing import biasfield
from pynet.preprocessing import register
from pynet.preprocessing import downsample
from pynet.preprocessing import padd
from pynet.preprocessing import scale


small_image = nibabel.Nifti1Image(
    downsample(image.get_data(), scale=4), image.affine)
processes = {
    "zscore_normalize": (zscore_normalize, {"mask": None}),
    "kde_normalize": (kde_normalize, {
        "mask": None, "modality": "T1w", "norm_value": 1}),
    "reorient2std": (reorient2std, {}),
    "register": (register, {
        "target": small_image, "cost": "corratio", "interp": "trilinear", "dof": 6}),
    "biasfield": (biasfield, {"nb_iterations": 3}),
    "downsample": (downsample, {"scale": 2}),
    "padd": (padd, {"shape": [256, 256, 256], "fill_value": 0}),
}

Test preprocessings¶

We now apply the preprocessing steps on the loaded image. Results are directly displayed in your browser at http://localhost:8097.

from pynet.plotting import Board

board = Board(port=8097, host="http://localhost", env="data-preprocessing")
for key, (fct, kwargs) in processes.items():
    print("Processing {0}...".format(key))
    if key in ("reorient2std", "biasfield", "register"):
        normalized = fct(small_image, **kwargs).get_data()
    else:
        normalized = fct(small_image.get_data(), **kwargs)
    if key in ("padd", "downsample", "register"):
        images = np.expand_dims(rescale(normalized, dynamic=(0, 255)), axis=0)
    else:
        images = np.asarray([rescale(small_image.get_data(), dynamic=(0, 255)),
                             rescale(normalized, dynamic=(0, 255))])
    images = images[..., images.shape[-1] // 2]
    images = np.expand_dims(images, axis=1)
    board.viewer.images(
        images, opts={"title": key, "caption": key}, win=key)
    print("Done.")
time.sleep(10)

Preprocessing pipeline¶

We now illustrate how we can use the Processor to preprocess the MRI images. Results are directly displayed in your browser at http://localhost:8097.

from pynet.preprocessing import Processor

board = Board(port=8097, host="http://localhost", env="data-preprocessing")
pipeline = Processor()
pipeline.register(reorient2std, apply_to="image")
pipeline.register(scale, scale=2, apply_to="image")
pipeline.register(biasfield, apply_to="image")
pipeline.register(register, target=target, apply_to="image")
pipeline.register(zscore_normalize, apply_to="array")
key = "pipeline"
normalized = pipeline(image)
print(image.shape, target.shape, normalized.shape)
images = np.asarray([rescale(target.get_data(), dynamic=(0, 255)),
                     rescale(normalized.get_data(), dynamic=(0, 255))])
images = images[..., images.shape[-1] // 2]
images = np.expand_dims(images, axis=1)
board.viewer.images(
    images, opts={"title": key, "caption": key}, win=key)

Total running time of the script: ( 0 minutes 0.000 seconds)

Gallery generated by Sphinx-Gallery