-
Notifications
You must be signed in to change notification settings - Fork 1
/
test_kreg_with_ref.py
230 lines (195 loc) · 9.69 KB
/
test_kreg_with_ref.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
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
import matplotlib.pyplot as plt
import numpy as np
import argparse
import os
import json
from pathlib import Path
import utils.vis
from utils import basic
import numpy as np
import medutils
def main(path, device=0, gif=False, overwrite=False):
# Specifiy experiment of iconik
exp_name = path
results_path = exp_name + "/rec_test"
phantom = True
config = basic.parse_config(exp_name + "/model_checkpoints/config.yml")
sub = config["subject_name"]
slice = config["slice"]
if "phantom" in sub:
if "0.33FS" in sub:
acc_factor = 3
elif "0.5FS" in sub:
acc_factor = 2
elif "1FS" in sub:
acc_factor = 1
else:
acc_factor = config["dataset"]["acc_factor"]
model = np.load(exp_name + '/rec_test/recon_best_network.npz')
# check if model is a kreg method
if "pretrained" not in config["model"]:
return
else:
pass
if os.path.exists(results_path + '/eval.txt') and not overwrite:
print("Not overwriting {}".format(results_path))
return
### grasp reference
grasp_path = config["data_root"]
grasp = np.load(grasp_path + "S" + str(config["subject_name"]) + "/grasp_reference_{}.npz".format(slice), allow_pickle=True)
### load reference NIK
nik_group_name = config["model"]["pretrained"]["pretrain_group"] + "_S" + str(config["subject_name"])
nik_exp_name = config["model"]["pretrained"]["pretrain_exp"] + "*" + "_slice" + str(config["slice"]) + "_R" + str(config['dataset']['acc_factor'])
nik_group_path = basic.find_subfolder(config["results_root"], nik_group_name)
nik_exp_path = basic.find_subfolder(nik_group_path, nik_exp_name)
nik_path = os.listdir(nik_exp_path)[0]
nik = np.load(nik_exp_path + "/" + nik_path + '/rec_test/recon_best_network.npz', allow_pickle=True)
# identify
## load data
img = {}
eval_dict ={}
img["model"] = model["recon"]
img["nik"] = nik["recon"]
img["inufft"] = grasp["R1"].item()["INUFFTnufft"] if "phantom" in sub else grasp["R{}".format(acc_factor)].item()["INUFFTnufft"]
img["xdgrasp4"] = grasp["R1"].item()["4MSgrasp"] if "phantom" in sub else grasp["R{}".format(acc_factor)].item()["4MSgrasp"]
img["nufft4"] = grasp["R1"].item()["4MSnufft"] if "phantom" in sub else grasp["R{}".format(acc_factor)].item()["4MSnufft"]
img["xdgrasp50"] = grasp["R1"].item()["50MSgrasp"] if "phantom" in sub else grasp["R{}".format(acc_factor)].item()["50MSgrasp"]
img["nufft50"] = grasp["R1"].item()["50MSnufft"] if "phantom" in sub else grasp["R{}".format(acc_factor)].item()["50MSnufft"]
if "phantom" in sub:
img["ref"] = model["ref"].transpose(3,4,2,0,1)
elif "knee" in sub:
img["ref"] = grasp["R1"].item()["INUFFTnufft"].repeat(100,1,1,1,1) # grasp["R{}".format(acc_factor)].item()["INUFFTnufft"].repeat(100,1,1,1,1)
elif "11_R0" in sub:
ref_path = config["data_root"]
ref = np.load(ref_path + "S11_gated" + "/grasp_reference_{}.npz".format(slice), allow_pickle=True)
img["ref"] = ref["R1"].item()["INUFFTnufft"].repeat(100,1,1,1,1)
else:
img["ref"] = grasp["R1"].item()["4MSgrasp"]
### Process images and calculate metrics
import utils.eval as eval
import utils.vis as vis
ech = 0
t = 10
img["ref"] = basic.torch2numpy(img["ref"])
img["ref"] = medutils.visualization.contrastStretching(img["ref"])
img["ref"] = medutils.visualization.normalize(img["ref"], max_int=1)
eval_str_xy, eval_str_xt, eval_str_yt = {},{},{}
for key in img.keys():
if "ref" not in key:
# expand images to GT motion states
img[key] = basic.torch2numpy(img[key])
img[key] = eval.create_hystereses(img[key], dim_axis=0)
img[key] = eval.postprocess(img[key], img["ref"])
if "knee" in sub or "phantom" in sub:
eval_dict[key] = eval.get_eval_metrics(img[key][:,ech, ...], img["ref"][:, ech, ...])
elif "11_R0" in sub:
eval_dict[key] = eval.get_eval_metrics(img[key][[0], ech, ...], img["ref"][[t], ech, ...]) # calculate metric only for temproal value
eval_str_xy[key] = eval.make_string_from_value_dict(eval_dict[key], default_keys=["psnr", "fsim"])
eval_str_xt[key] = eval.make_string_from_value_dict(eval_dict[key], default_keys=["psnr", "fsim_xt"])
eval_str_yt[key] = eval.make_string_from_value_dict(eval_dict[key], default_keys=["psnr", "fsim_yt"])
## ToDo: save comparison images
# NUFFT, GRASP4, GRASP5, NIK, ICoNIK
if "knee" in sub:
plot_order = ["inufft", "nik", "model"]
# elif "11_R0" in sub:
# plot_order = ["ref", "inufft", "nik", "model"]
else:
plot_order = ["ref", "nufft4", "xdgrasp4","xdgrasp50", "nik", "model"]
img_sat = {}
img_sat_max_list=[]
for key in plot_order:
img_sat[key] = medutils.visualization.contrastStretching(img[key][...].squeeze(), saturated_pixel=0.015)
img_sat_max_list.append(img_sat[key].max())
# Plot Spatial images
plot_metrics = ["psnr", "fsim"]
img_xy_list, img_max_list, title_list, eval_list = [], [], [],[]
for key, entry in img_sat.items():
img_xy_list.append(entry[t,...].squeeze())
title_list.append(key)
if key != "ref":
eval_list.append(eval_str_xy[key])
else:
eval_list.append("")
zoom_region = (105, 145, 50, 90) if "knee" in sub else (135,185,70,120)
eval_list = eval_list if "R0" not in sub else None
vis.plot_grid_from_lists([img_xy_list], [img_sat_max_list], [eval_list], [title_list],
zoom_region=None,
path=results_path + "/recon_comp_xy_t{}.jpg".format(t))
vis.plot_grid_from_lists([img_xy_list], [img_sat_max_list], [eval_list], [title_list],
zoom_region=None,
path=results_path + "/recon_comp_xy_t{}.eps".format(t))
vis.plot_grid_from_lists([img_xy_list], [img_sat_max_list], [eval_list], None,
zoom_region=zoom_region, crop_vert=45,
path=results_path + "/recon_comp_xy_t{}_zoom.jpg".format(t))
vis.plot_grid_from_lists([img_xy_list], [img_sat_max_list], [eval_list], None,
zoom_region=zoom_region, crop_vert=45,
path=results_path + "/recon_comp_xy_t{}_zoom.eps".format(t))
# Plot Spatial images
t=5
img_xy_list, title_list, eval_list = [], [],[]
for key, entry in img_sat.items():
# temp_img = medutils.visualization.contrastStretching(img[key][t,...].squeeze(), saturated_pixel=0.015)
img_xy_list.append(entry[t,...].squeeze())
# img_max_list.append(temp_img.max())
title_list.append(key)
if key != "ref":
eval_list.append(eval_str_xy[key])
else:
eval_list.append("")
eval_list = eval_list if "R0" not in sub else None
vis.plot_grid_from_lists([img_xy_list], [img_sat_max_list], [eval_list], [title_list],
path=results_path + "/recon_comp_xy_t{}_.jpg".format(t))
### Plot Temporal images
y = 160
eval_metrics = ["fsim_xt"]
plt.imshow(img["ref"][0, ..., :y, :].squeeze(), cmap="gray")
plt.title("cut-line")
plt.show()
img_yt_list, title_list, eval_list = [], [],[]
for key, entry in img_sat.items():
img_yt_list.append(entry[:50,...,y,:].squeeze())
title_list.append(key)
if key != "ref":
eval_list.append(eval_str_xt[key])
else:
eval_list.append("")
eval_list = eval_list if "R0" not in sub else None
vis.plot_grid_from_lists([img_yt_list],[img_sat_max_list], eval_list = None, title_list=[title_list], zoom_region=None,
path=results_path + "/recon_comp_xt_y{}.jpg".format(y))
### Plot Temporal images
x = 75
plt.imshow(img["ref"][0, ..., :, :x].squeeze(), cmap="gray")
plt.title("cut-line")
plt.savefig(results_path + "/recon_comp_yt_x{}_cutline.jpg".format(x))
plt.show()
img_yt_list, title_list, eval_list = [], [],[]
for key, entry in img_sat.items():
img_yt_list.append(entry[:50,...,:,x].squeeze())
title_list.append(key)
if key != "ref":
eval_list.append(eval_str_yt[key])
else:
eval_list.append("")
eval_list = eval_list if "R0" not in sub else None
vis.plot_grid_from_lists([img_yt_list], [img_sat_max_list], eval_list = None, title_list=[title_list],
path=results_path + "/recon_comp_yt_x{}.jpg".format(x))
if gif:
eval_str = eval_str_xy if "R0" not in sub else None
total_duration = 5
knav = np.linspace(0, total_duration, img_sat["ref"].shape[0])
for img_name, img_value in img_sat.items():
utils.vis.save_gif(img_value, #str=eval_str[img_name],
numbers_array=np.around(knav, decimals=2),
filename=results_path + "/dyn_recon_{}.gif".format(img_name),
intensity_factor=1, total_duration=total_duration)
with open(results_path + '/eval.txt', 'w') as f:
json.dump(eval_dict, f, indent=4, default=basic.float32_serializer)
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('-p', '--path', type=str, default='configs/config_abdominal.yml')
parser.add_argument('-g', '--gpu', type=int, default=0)
parser.add_argument('-o', '--overwrite', type=str, default="true")
args = parser.parse_args()
# Manually interpret the string as a boolean value
args.overwrite = args.overwrite.lower() == 'true'
main(path=args.path, device=args.gpu, overwrite=args.overwrite)