-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest.py
178 lines (159 loc) · 7.93 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
import argparse
from torch.autograd import Variable
from torch.utils.data import DataLoader
from net import Net
from dataset import *
import matplotlib.pyplot as plt
from metrics import *
import os
import time
import pdb
from torchvision.transforms import ToTensor
from PIL import Image
import torch.nn.functional as F
os.environ['KMP_DUPLICATE_LIB_OK'] = 'TRUE'
parser = argparse.ArgumentParser(description="PyTorch BasicIRSTD test")
parser.add_argument("--model_names", default=['ACM'], type=list,
help="model_name: 'ACM', 'ALCNet', 'DNANet', 'ISNet', 'UIUNet', 'RDIAN', 'ISTDU-Net', 'U-Net', 'RISTDnet'")#['ACM', 'ALCNet','DNANet', 'ISNet', 'RDIAN', 'ISTDU-Net']
parser.add_argument("--pth_dirs", default=['PRCV2024_old/ACM_230.pth.tar'], type=list, help="checkpoint dir, default=None or ['NUDT-SIRST/ACM_400.pth.tar','NUAA-SIRST/ACM_400.pth.tar']")
parser.add_argument("--dataset_dir", default='/home/public/', type=str, help="train_dataset_dir") #'NUAA-SIRST/ACM_200.pth.tar'
parser.add_argument("--dataset_names", default=['PRCV2024'], type=list,
help="dataset_name: 'NUAA-SIRST', 'NUDT-SIRST', 'IRSTD-1K', 'SIRST3', 'NUDT-SIRST-Sea'")#, 'NUDT-SIRST', 'IRSTD-1K'
parser.add_argument("--img_norm_cfg", default=None, type=dict,
help="specific a img_norm_cfg, default=None (using img_norm_cfg values of each dataset)")
parser.add_argument("--save_img", default=True, type=bool, help="save image of or not")
parser.add_argument("--save_img_dir", type=str, default='./results/', help="path of saved image")
parser.add_argument("--save_log", type=str, default='./log/', help="path of saved .pth")
parser.add_argument("--threshold", type=float, default=0.25)
def downsample_if_needed(img, size_limit=512):
"""如果图像尺寸超过限制，进行下采样"""
_,_,h, w = img.shape
if max(h, w) > size_limit:
scale_factor = size_limit / max(h, w)
new_h = int(h * scale_factor)
new_w = int(w * scale_factor)
img=F.interpolate(img, size=(new_h, new_w), mode='bilinear', align_corners=False)
#img = img.resize((new_w, new_h), resample=Image.BILINEAR)
return img, h,w
else:
return img, h,w
global opt
opt = parser.parse_args()
# def slice_tensor(tensor, slice_size=(256, 256), stride=128):
# """
# 对Tensor图像进行切片处理。
# """
# width, height = tensor.shape[-2:]
# slices = []
# for y in range(0, height, stride):
# for x in range(0, width, stride):
# end_x = min(x + slice_size[0], width)
# end_y = min(y + slice_size[1], height)
# slices.append(tensor[..., y:end_y, x:end_x])
# return slices
# def merge_tensor_slices(sliced_tensors, original_shape, slice_size=(256, 256), stride=128):
# """
# 合并切片后的Tensor预测结果。
# """
# merged_tensor = torch.zeros(original_shape, device=tensor.device)
# width, height = original_shape[-2:]
# num_cols = (width - slice_size[0]) // stride + 1
# num_rows = (height - slice_size[1]) // stride + 1
# for row in range(num_rows):
# for col in range(num_cols):
# start_x = col * stride
# start_y = row * stride
# end_x = min(start_x + slice_size[0], width)
# end_y = min(start_y + slice_size[1], height)
# merged_tensor[..., start_y:end_y, start_x:end_x] += sliced_tensors[row * num_cols + col]
# # 如果需要平均处理重叠区域，可以在这里添加相应逻辑
# # merged_tensor /= ... # 根据重叠次数进行归一化
# return merged_tensor
def test():
test_set = TestSetLoader(opt.dataset_dir, opt.train_dataset_name, opt.test_dataset_name, opt.img_norm_cfg)
test_loader = DataLoader(dataset=test_set, num_workers=1, batch_size=1, shuffle=False)
net = Net(model_name=opt.model_name, mode='test').cuda()
net.load_state_dict(torch.load(opt.pth_dir)['state_dict'])
net.eval()
eval_mIoU = mIoU()
eval_PD_FA = PD_FA()
tta=True
for idx_iter, (img, gt_mask, size, img_dir) in enumerate(test_loader):
#img, h,w = downsample_if_needed(img)
img = Variable(img).cuda()
pred = net.forward(img)
#pred=F.interpolate(pred, size=(h, w), mode='bilinear', align_corners=False)
#img_slices = slice_tensor(img, slice_size=(256, 256), stride=128)
#all_pred_slices = []
# for slice in img_slices:
# # 确保slice在GPU上进行推理
# slice = slice.cuda()
# pred_slice = net(slice.unsqueeze(0)) # 添加batch维度
# all_pred_slices.append(pred_slice.squeeze(0)) # 移除添加的batch维度
# 合并切片预测结果
#pred = merge_tensor_slices(all_pred_slices, img.shape)
#pred = pred[:,:,:size[0],:size[1]]
# if tta:
# #x,y,xy flips as TTA
# flips = [[-1],[-2],[-2,-1]]
# for f in flips:
# img = torch.flip(img,f)
# y_preds = net.forward(img)
# y_preds = torch.flip(y_preds,f)
# #y_pred = y_pred[:,:,:size[0],:size[1]]
# pred+=y_preds
# pred=pred/(1+len(flips))
pred=pred[:,:,:size[0],:size[1]]
#pdb.set_trace()
gt_mask = gt_mask[:,:,:size[0],:size[1]]
eval_mIoU.update((pred>opt.threshold).cpu(), gt_mask)
eval_PD_FA.update((pred[0,0,:,:]>opt.threshold).cpu(), gt_mask[0,0,:,:], size)
### save img
if opt.save_img == True:
img_save = transforms.ToPILImage()((pred[0,0,:,:]).cpu())
if not os.path.exists(opt.save_img_dir + opt.test_dataset_name + '/' + opt.model_name):
os.makedirs(opt.save_img_dir + opt.test_dataset_name + '/' + opt.model_name)
img_save.save(opt.save_img_dir + opt.test_dataset_name + '/' + opt.model_name + '/' + img_dir[0] + '.png')
results1 = eval_mIoU.get()
results2 = eval_PD_FA.get()
print("pixAcc, mIoU:\t" + str(results1))
print("PD, FA:\t" + str(results2))
opt.f.write("pixAcc, mIoU:\t" + str(results1) + '\n')
opt.f.write("PD, FA:\t" + str(results2) + '\n')
if __name__ == '__main__':
opt.f = open(opt.save_log + 'test_Sun_Jun__2_16_56_49_2024.txt','w')#'test_' + (time.ctime()).replace(' ', '_').replace(':', '_') + '.txt', 'w')
#pdb.set_trace()
if opt.pth_dirs == None:
for i in range(len(opt.model_names)):
opt.model_name = opt.model_names[i]
print(opt.model_name)
opt.f.write(opt.model_name + '_400.pth.tar' + '\n')
for dataset_name in opt.dataset_names:
opt.dataset_name = dataset_name
opt.train_dataset_name = opt.dataset_name
opt.test_dataset_name = opt.dataset_name
print(dataset_name)
opt.f.write(opt.dataset_name + '\n')
opt.pth_dir = opt.save_log + opt.dataset_name + '/' + opt.model_name + '_400.pth.tar'
test()
print('\n')
opt.f.write('\n')
opt.f.close()
else:
for model_name in opt.model_names:
for dataset_name in opt.dataset_names:
for pth_dir in opt.pth_dirs:
if dataset_name in pth_dir and model_name in pth_dir:
#pdb.set_trace()
opt.test_dataset_name = dataset_name
opt.model_name = model_name
opt.train_dataset_name = dataset_name #pth_dir.split('/')[0]
print(pth_dir)
opt.f.write(pth_dir)
print(opt.test_dataset_name)
opt.f.write(opt.test_dataset_name + '\n')
opt.pth_dir = opt.save_log + pth_dir
test()
print('\n')
opt.f.write('\n')
opt.f.close()