1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
|
"""
=====================================================
Tissue Classification using Diffusion MRI with DAM
=====================================================
This example demonstrates tissue classification of white matter (WM) and gray matter
(GM) from multi-shell diffusion MRI data using the Directional Average Maps (DAM)
proposed by :footcite:p:`Cheng2020`. DAM uses the diffusion properties of the tissue to
classify the voxels into WM and GM by fitting a polynomial model to the diffusion
signal. The process involves preprocessing steps including skull-stripping with
median otsu, denoising with Patch2Self, and then perform tissue classification.
Let's start by loading the necessary modules:
"""
import matplotlib.pyplot as plt
from dipy.core.gradients import gradient_table
from dipy.data import get_fnames
from dipy.denoise.patch2self import patch2self
from dipy.io.gradients import read_bvals_bvecs
from dipy.io.image import load_nifti
from dipy.segment.mask import median_otsu
from dipy.segment.tissue import dam_classifier
from dipy.viz.plotting import image_mosaic
###############################################################################
# First we fetch the diffusion image, bvalues and bvectors from the cfin dataset.
fraw, fbval, fbvec, t1_fname = get_fnames(name="cfin_multib")
data, affine = load_nifti(fraw)
bvals, bvecs = read_bvals_bvecs(fbval, fbvec)
gtab = gradient_table(bvals, bvecs=bvecs)
###############################################################################
# After loading the diffusion data, we can apply the median_otsu algorithm to
# skull-strip the data and obtain a binary mask. We can then use the mask to
# denoise the data using the Patch2Self algorithm.
b0_mask, mask = median_otsu(data, median_radius=2, numpass=1, vol_idx=[0, 1])
denoised_arr = patch2self(b0_mask, bvals=bvals, b0_denoising=False)
###############################################################################
# Now we can use the DAM classifier to classify the voxels into WM and GM.
# The DAM classifier requires the denoised data, the bvalues, and the mask.
# The DAM classifier returns the WM and GM masks.
# It is important to note that the DAM classifier is a threshold-based classifier
# and the threshold values can be adjusted based on the data. The `wm_threshold`
# parameter controls the sensitivity of the classifier.
# For data like HCP, threshold of 0.5 proves to be a good choice. For data like
# cfin, higher threshold values like 0.7 or 0.8 are more suitable.
wm_mask, gm_mask = dam_classifier(denoised_arr, bvals, wm_threshold=0.7)
###############################################################################
# Now we can visualize the WM and GM masks.
images = [
data[:, :, data.shape[2] // 2, 0], # DWI (b0)
wm_mask[:, :, data.shape[2] // 2], # White Matter Mask
gm_mask[:, :, data.shape[2] // 2], # Grey Matter Mask
]
ax_labels = ["DWI (b0)", "White Matter Mask", "Grey Matter Mask"]
ax_kwargs = [{"cmap": "gray"} for _ in images]
fig, ax = image_mosaic(
images, ax_labels=ax_labels, ax_kwargs=ax_kwargs, figsize=(10, 5)
)
plt.show()
###############################################################################
# .. rst-class:: centered small fst-italic fw-semibold
#
# Original B0 image (left), White matter (center) and gray matter (right) are
# binary masks as obtained from DAM.
#
#
# References
# ----------
#
# .. footbibliography::
|
