-
Notifications
You must be signed in to change notification settings - Fork 22
/
run_CT_recon.py
165 lines (139 loc) · 6.2 KB
/
run_CT_recon.py
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
159
160
161
162
163
164
165
import torch
# from losses import get_optimizer
from models.ema import ExponentialMovingAverage
import numpy as np
import controllable_generation
from utils import restore_checkpoint, show_samples_gray, clear, clear_color, \
lambda_schedule_const, lambda_schedule_linear
from pathlib import Path
from models import utils as mutils
from models import ncsnpp
from sde_lib import VESDE
from sampling import (ReverseDiffusionPredictor,
LangevinCorrector)
import datasets
import time
# for radon
from physics.ct import CT
import matplotlib.pyplot as plt
###############################################
# Configurations
###############################################
solver = 'MCG'
config_name = 'AAPM_256_ncsnpp_continuous'
sde = 'VESDE'
num_scales = 2000
ckpt_num = 185
N = num_scales
root = './samples/CT'
# Parameters for the inverse problem
sparsity = 6
num_proj = 180 // sparsity # 180 / 6 = 30
det_spacing = 1.0
size = 256
det_count = int((size * (2*torch.ones(1)).sqrt()).ceil()) # ceil(size * \sqrt{2})
schedule = 'linear'
start_lamb = 1.0
end_lamb = 0.6
num_posterior_sample = 1
if schedule == 'const':
lamb_schedule = lambda_schedule_const(lamb=start_lamb)
elif schedule == 'linear':
lamb_schedule = lambda_schedule_linear(start_lamb=start_lamb, end_lamb=end_lamb)
else:
NotImplementedError(f"Given schedule {schedule} not implemented yet!")
freq = 1
if sde.lower() == 'vesde':
from configs.ve import AAPM_256_ncsnpp_continuous as configs
ckpt_filename = f"./checkpoints/{config_name}/checkpoint_{ckpt_num}.pth"
config = configs.get_config()
config.model.num_scales = N
sde = VESDE(sigma_min=config.model.sigma_min, sigma_max=config.model.sigma_max, N=config.model.num_scales)
sde.N = N
sampling_eps = 1e-5
batch_size = 1
config.training.batch_size = batch_size
predictor = ReverseDiffusionPredictor
corrector = LangevinCorrector
probability_flow = False
snr = 0.16
n_steps = 1
batch_size = 1
config.training.batch_size = batch_size
config.eval.batch_size = batch_size
random_seed = 0
sigmas = mutils.get_sigmas(config)
scaler = datasets.get_data_scaler(config)
inverse_scaler = datasets.get_data_inverse_scaler(config)
score_model = mutils.create_model(config)
# optimizer = get_optimizer(config, score_model.parameters())
ema = ExponentialMovingAverage(score_model.parameters(),
decay=config.model.ema_rate)
# state = dict(step=0, optimizer=optimizer,
# model=score_model, ema=ema)
state = dict(step=0, model=score_model, ema=ema)
state = restore_checkpoint(ckpt_filename, state, config.device, skip_sigma=True, skip_optimizer=True)
ema.copy_to(score_model.parameters())
idx = 100
filename = Path(root) / (str(idx).zfill(4) + '.npy')
# Specify save directory for saving generated samples
save_root = Path(f'./results/SV-CT/m{180/sparsity}/{idx}/{solver}')
save_root.mkdir(parents=True, exist_ok=True)
irl_types = ['input', 'recon', 'label']
for t in irl_types:
save_root_f = save_root / t
save_root_f.mkdir(parents=True, exist_ok=True)
# Read data
img = torch.from_numpy(np.load(filename))
h, w = img.shape
img = img.view(1, 1, h, w)
img = img.to(config.device)
plt.imsave(save_root / 'label' / f'{str(idx).zfill(4)}.png', clear(img), cmap='gray')
plt.imsave(save_root / 'label' / f'{str(idx).zfill(4)}_clip.png', np.clip(clear(img), 0.1, 1.0), cmap='gray')
# full
angles = np.linspace(0, np.pi, 180, endpoint=False)
radon = CT(img_width=size, radon_view=num_proj, circle=False, device=config.device)
radon_all = CT(img_width=size, radon_view=180, circle=False, device=config.device)
mask = torch.zeros([batch_size, 1, det_count, 180]).to(config.device)
mask[..., ::sparsity] = 1
# Dimension Reducing (DR)
sinogram = radon.A(img)
# Dimension Preserving (DP)
sinogram_full = radon_all.A(img) * mask
# FBP
fbp = radon.A_dagger(sinogram)
plt.imsave(str(save_root / 'input' / f'FBP.png'), clear(fbp), cmap='gray')
if solver == 'MCG':
pc_MCG = controllable_generation.get_pc_radon_MCG(sde,
predictor, corrector,
inverse_scaler,
snr=snr,
n_steps=n_steps,
probability_flow=probability_flow,
continuous=config.training.continuous,
denoise=True,
radon=radon,
radon_all=radon_all,
weight=0.1,
save_progress=False,
save_root=save_root,
lamb_schedule=lamb_schedule,
mask=mask)
x = pc_MCG(score_model, scaler(img), measurement=sinogram)
elif solver == 'song':
pc_song = controllable_generation.get_pc_radon_song(sde,
predictor, corrector,
inverse_scaler,
snr=snr,
n_steps=n_steps,
probability_flow=probability_flow,
continuous=config.training.continuous,
save_progress=True,
save_root=save_root,
denoise=True,
radon=radon_all,
lamb=0.7)
x = pc_song(score_model, scaler(img), mask, measurement=sinogram_full)
# Recon
plt.imsave(str(save_root / 'recon' / f'{str(idx).zfill(4)}.png'), clear(x), cmap='gray')
plt.imsave(str(save_root / 'recon' / f'{str(idx).zfill(4)}_clip.png'), np.clip(clear(x), 0.1, 1.0), cmap='gray')