-
Notifications
You must be signed in to change notification settings - Fork 0
/
predict.py
163 lines (131 loc) · 5.8 KB
/
predict.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
import argparse
import logging
import os
import numpy as np
import torch
import torch.nn.functional as F
from PIL import Image
from torchvision import transforms
from utils.data_loading import BasicDataset
from unet import UNet, UnetResnet50, hrnet48, Unet_p1, hrnet48_p1, UNet_fp16, UNet_fp4
from utils.utils import plot_img_and_mask
import cv2
import time
def predict_img(net,
full_img,
device,
scale_factor=1,
out_threshold=0.5):
net.eval()
img = BasicDataset.preprocess(full_img, scale_factor, is_mask=False,is_transforms=False)
img = img.transpose((2, 0, 1))
img = torch.from_numpy(img)
img = img.unsqueeze(0)
img = img.to(device=device, dtype=torch.float32)
# print('input img shape: ',img.shape)
with torch.no_grad():
output = net(img)
# print('output img shape: ',output.shape)
if net.n_classes > 1:
probs = F.softmax(output, dim=1)[0]
else:
probs = torch.sigmoid(output)[0]
# print('F.softmax(output, dim=1).shape: ',F.softmax(output, dim=1).shape)
# print('probs.shape: ',probs.shape)
tf = transforms.Compose([
transforms.ToPILImage(),
transforms.Resize((full_img.shape[0], full_img.shape[1])),
transforms.ToTensor()
])
full_mask = tf(probs.cpu()).squeeze()
# print('full_mask.shape: ',full_mask.shape)
if net.n_classes == 1:
return (full_mask > out_threshold).numpy()
else:
return F.one_hot(full_mask.argmax(dim=0), net.n_classes).permute(2, 0, 1).numpy()
def get_args():
parser = argparse.ArgumentParser(description='Predict masks from input images')
parser.add_argument('--model', '-m', default='MODEL.pth', metavar='FILE',
help='Specify the file in which the model is stored')
parser.add_argument('--input', '-i', metavar='INPUT', nargs='+', help='Filenames of input images', required=True)
parser.add_argument('--output', '-o', metavar='INPUT', nargs='+', help='Filenames of output images')
parser.add_argument('--viz', '-v', action='store_true',
help='Visualize the images as they are processed')
parser.add_argument('--no-save', '-n', action='store_true', help='Do not save the output masks')
parser.add_argument('--mask-threshold', '-t', type=float, default=0.5,
help='Minimum probability value to consider a mask pixel white')
parser.add_argument('--scale', '-s', type=float, default=1, # 0.5
help='Scale factor for the input images')
return parser.parse_args()
def get_output_filenames(args):
def _generate_name(fn):
split = os.path.splitext(fn)
return f'{split[0]}_OUT{split[1]}'
return args.output or list(map(_generate_name, args.input))
def mask_to_image(mask: np.ndarray):
if mask.ndim == 2:
# print('mask.ndim = 2')
return Image.fromarray((mask * 255).astype(np.uint8))
elif mask.ndim == 3:
# print('mask.ndim = 3')
channel=mask.shape[0]
if channel==3:
backgrand=Image.fromarray(np.uint8(mask[0]*0))
class1=Image.fromarray(np.uint8(mask[1]*255))
class2=Image.fromarray(np.uint8(mask[2]*255))
result_img=Image.merge('RGB',(class2,class1,backgrand))
elif channel==2:
result_img=Image.fromarray(np.uint8(mask[1]*255))
return result_img
# return Image.fromarray((np.argmax(mask, axis=0) * 255 / mask.shape[0]).astype(np.uint8))
if __name__ == '__main__':
# print()
# print('*************************predict***********************************')
# 统计用时
start_time=time.perf_counter()
args = get_args()
in_files = args.input
out_files = get_output_filenames(args)
net = UNet(n_channels=3, n_classes=2)
# net = UNet(n_channels=3, n_classes=2, bilinear=True)
# net = UnetResnet50(n_channels=3, n_classes=2)
# net =hrnet48(n_channels=3, n_classes=2)
# net =Unet_p1(n_channels=3, n_classes=2)
# net = hrnet48_p1(n_channels=3, n_classes=2)
# net = UNet_fp4(n_channels=3, n_classes=2)
# net = UNet_fp16(n_channels=3, n_classes=2)
# device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
device = torch.device('cpu')
logging.info(f'Loading model {args.model}')
logging.info(f'Using device {device}')
net.to(device=device)
net.load_state_dict(torch.load(args.model, map_location=device))
logging.info('Model loaded!')
num_img=0
for i, filename in enumerate(in_files):
num_img+=1
# print('filename: ',filename)
logging.info(f'\nPredicting image {filename} ...')
# img = Image.open(filename)
img=cv2.imread(filename)
# print('img: ',img)
mask = predict_img(net=net,
full_img=img,
scale_factor=args.scale,
out_threshold=args.mask_threshold,
device=device)
# print('mask_pre type: ',type(mask))
# print('mask_pre shape: ',mask.shape)
# print('mask_pre: ',mask)
if not args.no_save:
out_filename = out_files[i]
result = mask_to_image(mask)
result.save(out_filename)
logging.info(f'Mask saved to {out_filename}')
if args.viz:
logging.info(f'Visualizing results for image {filename}, close to continue...')
plot_img_and_mask(img, mask)
# 统计用时
end_time=time.perf_counter()
print('总共用时：',end_time-start_time)
print('平均用时：',(end_time-start_time)/num_img)