/
exp_knee_pf.m
161 lines (102 loc) · 3.29 KB
/
exp_knee_pf.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
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
%% Partial Fourier
clc
clear
close all
setPath
%% Get Kspace
load knee
ksp = ksp/max(max(max(abs(ifft2c(ksp)))));
FOV = size(ksp);
[sx, sy, nc] = size(ksp);
%% Undersampling
frac = 7/16; % partial fourier fraction
ncalib = 20;
mask = ones(sx, sy);
mask(:, 1:round(FOV(2)*frac)) = 0;
mask(:, (sy-ncalib)/2:(sy+ncalib)/2) = 1;
figure, imshow3( mask);
mask = repmat( mask, [1,1,FOV(3:end)]);
y = ksp.*mask;
mask = y ~= 0;
figure, imshow3( log(abs(y)) );
titlef('Log Magnitude of Kspace');
%% Generate Maps
ksize = [6, 6];
[maps, weights] = ecalib(y, ncalib, ksize);
figure, imshow3(abs(maps .* repmat( weights,1,1,nc)));
titlef('Sensitivity Maps');
%% Create linear operators
C = Identity;
S = ESPIRiT(maps, weights);
F = p2DFT(mask,[sx, sy, nc]);
M = Identity;
P = Identity;
%% Obtain low frequency phase estimate
win = repmat(hanning(ncalib) * hanning(ncalib)', [1, 1, nc]);
p_est = angle(S' * (F' * zpad(crop(ksp, ncalib, ncalib, nc) .* win, sx, sy, nc)));
figure, imshow3(p_est)
mapsp = maps .* repmat(exp(1j*p_est), [1, 1, nc]);
Sp = ESPIRiT(mapsp, weights);
%% Create proximal operators
lambdam = 0.003;
lambdap = 0.003;
Pm = wave_thresh('db4', 3, lambdam);
Pp = wave_thresh('db4', 3, lambdap);
%% Get Fully-sampled Image
x = S' * (ifft2c(ksp));
figure, imshowf(abs(x), [0, 1.0])
figure, imshowf(abs(abs(x) - abs(x)), [0, 0.1])
figure, imshowf(angle(x) .* (abs(x) > 0.1), [-pi, pi])
%% Zero-filled recon
x_sub = S' * ( F' * y );
figure, imshowf(abs(x_sub), [0, 1.0])
figure, imshowf(abs(abs(x_sub) - abs(x)), [0, 0.1])
figure, imshowf(angle(x_sub) .* (abs(x) > 0.1), [-pi, pi])
%% Bydder et al.
maxiter = 1000;
doplot = 1;
dohogwild = 1;
x_im = imhomodyne(y, Sp, F, Pm, Pp, maxiter, dohogwild, doplot);
x_im = x_im .* sqrt(weights);
disp(psnr(abs(x), abs(x_im)))
figure, imshow(abs(x_im), [0, 1.0])
figure, imshow(abs(abs(x_im) - abs(x)), [0, 0.3])
figure, imshow(angle(x_im .* exp(1j * p_est)) .* (abs(x) > 0.1), [-pi, pi])
%% Get phase wraps
ncycles = 16;
[m0, p0, W] = pfinit(y, S, F, ncycles);
%% Proposed method without phase cycling
niter = 100;
ninneriter = 10;
doplot = 1;
dohogwild = 1;
[mn, pn] = mprecon(y, F, S, C, M, P, Pm, Pp, m0, p0, {}, niter, ninneriter, dohogwild, doplot);
mn = mn .* sqrt(weights);
disp(psnr(abs(x), abs(mn)))
%%
figure, imshowf(abs(mn), [0, 1.0])
figure, imshowf(abs(mn - abs(x)), [0, 0.3])
figure, imshowf(pn .* (abs(x) > 0.1), [-pi, pi])
%% Proposed method with phase cycling
niter = 100;
ninneriter = 10;
doplot = 1;
dohogwild = 1;
[m, p] = mprecon(y, F, S, C, M, P, Pm, Pp, m0, p0, W, niter, ninneriter, dohogwild, doplot);
m = m .* sqrt(weights);
disp(psnr(abs(x), abs(m)))
figure, imshow(abs(m), [0, 1.0])
figure, imshow(abs(abs(m) - abs(x)), [0, 0.3])
figure, imshow(p .* (abs(x) > 0.1), [-pi, pi])
%% Zhao et al. Separate magnitude and phase reconstruction
% Requires irt from Jeff Fessler.
% Please run setup.m in the toolbox first.
lambda_m = 0.1; % regularization parameter for magnitude
lambda_p = 0.1; % regularization parameter for phase (rg2/rg4)
im_mask = weights > 0.1;
maps = maps .* im_mask;
proxg_m = wave_thresh('db4', 3, lambda_m);
y = ksp(mask == 1);
samp = mask(:, :, 1) == 1; % change to logical for irt.
[mi, xi] = separate_mag_phase_recon(y, samp, maps, im_mask, proxg_m, lambda_p);
disp(psnr(abs(x) / max(abs(x(:))), abs(mi) / max(abs(mi(:)))))