-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathtest.py
190 lines (149 loc) · 7.72 KB
/
test.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
import os
import time
import pickle
import logging
import cv2
import numpy as np
from metrics.metric_group import MetricGroup
cv2.setNumThreads(0)
cv2.ocl.setUseOpenCL(False)
def test(model, data_loader, device, epoch, args):
model.eval()
dataset = data_loader.dataset
metric_group = MetricGroup(accuracyD=args.test_config["METRIC"]["accuracyD"],
accuracyI=args.test_config["METRIC"]["accuracyI"],
accuracyC=args.test_config["METRIC"]["accuracyC"],
ECED=args.test_config["METRIC"]["ECED"],
ECEI=args.test_config["METRIC"]["ECEI"],
SCED=args.test_config["METRIC"]["SCED"],
SCEI=args.test_config["METRIC"]["SCEI"],
q=args.test_config["METRIC"]["q"],
binary=args.data_config["num_classes"] == 2,
num_bins=args.test_config["METRIC"]["num_bins"],
num_classes=args.data_config["num_classes"],
ignore_index=args.data_config["num_classes"])
end = time.time()
for iter, (image, label, image_file) in enumerate(data_loader):
start = time.time()
toprint = f"Epoch: [{epoch}|{args.schedule_config['train_epochs']}], "
toprint += f"Iter: [{iter}|{len(data_loader)}], "
toprint += f"Data Time: {(start - end):.6f}, "
prob = model.multi_scale_predict(image,
device,
args.data_config["num_classes"],
args.data_config["crop_size"],
args.test_config["INF"]["flip"],
args.test_config["INF"]["ratios"],
args.test_config["INF"]["stride_rate"])
mid = time.time()
toprint += f"Batch Time: {(mid - start):.6f}, "
metric_group.add(prob, label, image_file)
if args.test_config["ITER"]["save_pred"]:
save_pred(prob, label, image_file, dataset.color_map, args)
end = time.time()
toprint += f"Metric Time: {(end - mid):.6f}"
if iter % args.test_config["ITER"]["log_iters"] == 0:
logging.info(toprint)
toprint = "\n"
results = metric_group.value()
for key, value in results.items():
toprint += f"{key}: {value:.2f}\n"
toprint = toprint[:-2]
logging.info(toprint)
with open(os.path.join(args.output_dir, "metric_group.pkl"), "wb") as pkl:
pickle.dump(metric_group, pkl)
def test_medical(model, data_loader, device, epoch, args):
model.eval()
val_cases = [False] * (args.data_config["num_cases"] + 1)
metric_groups = {}
dataset = data_loader.dataset
for case in range(args.data_config["num_cases"] + 1):
metric_groups[case] = MetricGroup(accuracyD=args.test_config["METRIC"]["accuracyD"],
accuracyI=args.test_config["METRIC"]["accuracyI"],
accuracyC=args.test_config["METRIC"]["accuracyC"],
ECED=args.test_config["METRIC"]["ECED"],
ECEI=args.test_config["METRIC"]["ECEI"],
SCED=args.test_config["METRIC"]["SCED"],
SCEI=args.test_config["METRIC"]["SCEI"],
q=args.test_config["METRIC"]["q"],
binary=args.data_config["num_classes"] == 2,
num_bins=args.test_config["METRIC"]["num_bins"],
num_classes=args.data_config["num_classes"],
ignore_index=args.data_config["num_classes"])
end = time.time()
for iter, (image, label, image_file) in enumerate(data_loader):
start = time.time()
toprint = f"Epoch: [{epoch}|{args.schedule_config['train_epochs']}], "
toprint += f"Iter: [{iter}|{len(data_loader)}], "
toprint += f"Data Time: {(start - end):.6f}, "
prob = model.multi_scale_predict(image,
device,
args.data_config["num_classes"],
args.data_config["crop_size"],
args.test_config["INF"]["flip"],
args.test_config["INF"]["ratios"],
args.test_config["INF"]["stride_rate"])
mid = time.time()
toprint += f"Batch Time: {(mid - start):.6f}, "
if "qubiq" in args.data_config["dataset"]:
ignore = label == args.data_config["num_classes"]
label = (label >= (args.data_config["num_raters"] // 2 + 1)).long()
label[ignore] = args.data_config["num_classes"]
label = label.long()
for i in range(image.shape[0]):
if args.data_config["dataset"] in ["lits", "kits"]:
case = int(image_file[i].split("/")[-1].split("_")[0])
elif "qubiq" in args.data_config["dataset"]:
case = int(image_file[i].split("/")[-2][-2:])
else:
raise NotImplementedError
val_cases[case] = True
prob_i = prob[i, :, :, :].unsqueeze(0)
label_i = label[i, :, :].unsqueeze(0)
metric_groups[case].add(prob_i, label_i, image_file)
if args.test_config["ITER"]["save_pred"]:
save_pred(prob, label, image_file, dataset.color_map, args)
end = time.time()
toprint += f"Metric Time: {(end - mid):.6f}"
if iter % args.test_config["log_iters"] == 0:
logging.info(toprint)
results = {}
for i, case in enumerate(val_cases):
if case:
results_case = metric_groups[i].value()
for key, value in results_case.items():
if key not in results:
results[key] = value
else:
results[key] += value
for key in results:
results[key] /= sum(val_cases)
toprint = ""
for key, value in results.items():
toprint += f"{key}: {value:.2f}\n"
toprint = toprint[:-2]
logging.info(toprint)
with open(os.path.join(args.output_dir, f"metric_groups.pkl"), "wb") as pkl:
pickle.dump(metric_groups, pkl)
def save_pred(prob, label, image_file, color_map, args):
pred_dir = os.path.join(args.output_dir, "pred")
os.makedirs(pred_dir, exist_ok=True)
pred = prob.argmax(1)
pred[label == args.data_config["num_classes"]] = args.data_config["num_classes"]
for i in range(label.shape[0]):
pred_i = pred[i, :, :].cpu().numpy()
label_i = label[i, :, :].cpu().numpy()
pred_rgb = np.zeros((label.shape[1], label.shape[2], 3))
label_rgb = np.zeros((label.shape[1], label.shape[2], 3))
for j in range(len(color_map)):
pred_rgb[:, :, 2][pred_i == j] = color_map[j][0]
pred_rgb[:, :, 1][pred_i == j] = color_map[j][1]
pred_rgb[:, :, 0][pred_i == j] = color_map[j][2]
label_rgb[:, :, 2][label_i == j] = color_map[j][0]
label_rgb[:, :, 1][label_i == j] = color_map[j][1]
label_rgb[:, :, 0][label_i == j] = color_map[j][2]
label_file = os.path.join(pred_dir, image_file[i].split("/")[-1])
label_file = label_file.replace(".jpg", ".png")
pred_file = label_file.replace(".png", "_pred.png")
cv2.imwrite(pred_file, pred_rgb)
cv2.imwrite(label_file, label_rgb)