-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathps_LST_lesfill_core.m
129 lines (111 loc) · 3.42 KB
/
ps_LST_lesfill_core.m
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
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
function filled = ps_LST_lesfill_core(img, prob, seg, indx_brain)
%UNTITLED Summary of this function goes here
% Detailed explanation goes here
% Correct seg, if image is not T1
t1 = 1; t2 = 0; f2 = 0;
%if ~((mean(img(seg == 1)) < mean(img(seg == 2))) && (mean(img(seg == 2)) < mean(img(seg == 3))))
if mean(img(prob > 0)) > mean(img(seg == 3))
t1 = 0;
f2 = 1;
y = img(seg == 1);
csf_mean = mean(y);
csf_sd = std(y);
csf_ci = [csf_mean - 1.5 * csf_sd, csf_mean + 1.5 * csf_sd];
seg_new = seg;
seg_new(img > csf_ci(1) & img < csf_ci(2) & seg > 0) = 1;
c_tmp = ps_count(seg_new(seg_new > 0));
if any(c_tmp(2,:) ./ sum(c_tmp(2,:)) > .8)
f2 = 0;
t2 = 1;
seg_new = seg;
end
seg = seg_new;
end
% expand lesions if surrounding voxels are not CSF
% expand lesions if surrounding voxels are not CSF
binary = 1 .* (prob > 0);
csf = 1 .* (seg == 1);
indx_img = 0 .* img; indx_img(indx_brain) = indx_brain;
binary2 = binary;
indx_tmp = find(binary > 0);
nh_seg = getNeighborhood2(seg, indx_tmp, 1);
nh_indx = getNeighborhood2(indx_img, indx_tmp, 1);
tmp = nh_indx .* (nh_seg > 1);
binary(tmp(tmp > 0)) = 1;
% expand CSF
indx_tmp = find(seg == 1);
nh_seg = getNeighborhood2(seg, indx_tmp, 1);
nh_indx = getNeighborhood2(indx_img, indx_tmp, 1);
tmp = nh_indx .* (nh_seg > 1);
csf(tmp(tmp > 0)) = 1;
% Grow from the outside
indx_nowm = indx_brain(binary(indx_brain) > 0 | csf(indx_brain) > 0);
filled = img; filled(indx_nowm) = 0;
indx_tmp = find(seg > 0 & filled > 0);
nh = getNeighborhood2(filled, indx_tmp, 1);
filled(indx_tmp(sum(nh > 0) < 3)) = 0;
binary(indx_tmp(sum(nh > 0) < 3)) = 1;
y = img(seg == 3);
gm_mean = mean(y);
gm_sd = std(y);
gm_ci = [gm_mean - 2 * gm_sd, gm_mean + 2 * gm_sd];
y = img(seg == 3 & filled > 0);
wm_mean = mean(y);
wm_sd = std(y);
wm_ci = [wm_mean - 2 * wm_sd, wm_mean + 2 * wm_sd];
y = img(seg == 1 & prob == 0);
csf_mean = mean(y);
csf_sd = std(y);
csf_ci = [csf_mean - 2 * csf_sd, csf_mean + 2 * csf_sd];
counter = 0; st = 0;
while ~st
counter = counter + 1;
indx_tmp = find(binary > 0);
%numel(indx_tmp)
if numel(indx_tmp) > 0
nh = getNeighborhood2(filled, indx_tmp, 1);
if t1
nh(nh < csf_ci(2)) = 0;
else
if f2
nh(nh < gm_ci(1) | nh > gm_ci(2)) = 0;
else
nh(nh < wm_ci(1) | nh > wm_ci(2)) = 0;
%nh(nh > csf_ci(1)) = 0;
end
end
nm = sum(nh) ./ sum(nh ~= 0);
nm(isnan(nm)) = 0;
if all(nh(:) == 0) | counter > 200
st = 1;
else
filled(indx_tmp) = nm;% + normrnd(0, nanmean(nm) * .001, 1, numel(nm));
%binary(indx_tmp(~isnan(nm))) = 0;
binary(indx_tmp(sum(nh) ~= 0)) = 0;
end
else
st = 1;
end
end
if any(binary(:) > 0)
csf(binary > 0) = 1;
end
indx_tmp = find(binary2 > 0);
nh = getNeighborhood2(filled, indx_tmp, 1);
tmp = bsxfun(@times, [filled(indx_tmp)'; nh], [1/3; repmat(1/9, 6, 1)]);
tmp(isnan(tmp)) = 0;
filled(indx_tmp) = sum(tmp);
filled(csf == 1 & binary2 == 0) = img(csf == 1 & binary2 == 0);
indx_tmp = find(seg > 0 & filled == 0);
if numel(indx_tmp) > 0
st = 0;
while ~st
nh = getNeighborhood2(filled, indx_tmp, 1);
filled(indx_tmp) = sum(nh) ./ sum(nh > 0);
indx_tmp = find(seg > 0 & filled == 0);
if isempty(indx_tmp)
st = 1;
end
end
end
end