-
Notifications
You must be signed in to change notification settings - Fork 7
/
normalize_patches.py
176 lines (149 loc) · 5.44 KB
/
normalize_patches.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
# +
import os
import torch
from torch import nn
from torch.autograd import Variable
import torchvision
import torchvision.datasets as dsets
import torchvision.transforms as transforms
import utils
from arch import define_Gen, define_Dis
import kornia
import pandas as pd
import warnings
import torch.nn.functional as F
import numpy as np
import json
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import argparse
from PIL import Image
import random
# -
warnings.filterwarnings('ignore')
class Arguments(object):
def __init__(self, dictionary):
"""Constructor"""
for key in dictionary:
setattr(self, key, dictionary[key])
# +
args = {
'epochs': 10,
'decay_epoch': 9,
'batch_size': 1,
'lr': 0.0002,
'load_height': 128,
'load_width': 128,
'gpu_ids': '0',
'crop_height': 128,
'crop_width': 128,
'alpha': 10, # Cyc loss
'beta': 5, # Scyc loss
'gamma': 5, # Dssim loss
'delta': 0.5, # Identity
'training': True,
'testing': True,
'results_dir': '/project/DSone/as3ek/data/ganstain/CCHMC_vsi_svs/results/',
'dataset_dir': '/project/DSone/as3ek/data/ganstain/CCHMC_vsi_svs/',
'checkpoint_dir': '/project/DSone/as3ek/data/ganstain/CCHMC_vsi_svs/checkpoint/',
'norm': 'batch',
'use_dropout': False,
'ngf': 64,
'ndf': 64,
'gen_net': 'unet_128',
'dis_net': 'n_layers',
'self_attn': True,
'spectral': True,
'log_freq': 50,
'custom_tag': '',
'gen_samples': False,
'specific_samples': False
}
args = Arguments(args)
# SOURCE AND TARGET FOLDERS
source_path = '/project/DSone/as3ek/data/patches/1000/classification/cinn_celiac__normal/valid/celiac/'
target_path = '/project/DSone/as3ek/data/patches/1000/gan_normalized/cinn_celiac__normal/valid/celiac/'
train_valid_split = 1
size = 256
one_direction = False # If this is false. b -> a -> b will happen. Edit code for otherwise.
gen_name = 'Gba' # Gba to generate b given a, i.e., a -> b
folder_to_folder = True
if not os.path.exists(target_path):
os.makedirs(target_path)
tag1 = 'noattn'
if args.self_attn:
tag1 = 'attn'
tag2 = 'nospec'
if args.spectral:
tag2 = 'spectral'
# Generate paths for checkpoint and results
args.identifier = str(args.gen_net) + '_' + str(args.dis_net) + '_' \
+ str(args.lr) + '_' + args.norm + '_' + tag1 + '_' + tag2 + '_' + str(args.batch_size) + '_' \
+ str(args.load_height) + '_coefs_' + str(args.alpha) + '_' + str(args.beta) + '_' + str(args.gamma) + '_'\
+ str(args.delta) + '_' + args.custom_tag
args.checkpoint_path = args.checkpoint_dir + args.identifier
args.results_path = args.results_dir + args.identifier
args.gpu_ids = []
for i in range(torch.cuda.device_count()):
args.gpu_ids.append(i)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# -
if one_direction:
G = define_Gen(input_nc=3, output_nc=3, ngf=args.ngf, netG=args.gen_net, norm=args.norm,
use_dropout= args.use_dropout, gpu_ids=args.gpu_ids, self_attn=args.self_attn, spectral = args.spectral)
else:
Gab = define_Gen(input_nc=3, output_nc=3, ngf=args.ngf, netG=args.gen_net, norm=args.norm,
use_dropout= args.use_dropout, gpu_ids=args.gpu_ids, self_attn=args.self_attn, spectral = args.spectral)
Gba = define_Gen(input_nc=3, output_nc=3, ngf=args.ngf, netG=args.gen_net, norm=args.norm,
use_dropout= args.use_dropout, gpu_ids=args.gpu_ids, self_attn=args.self_attn, spectral = args.spectral)
ckpt = utils.load_checkpoint('%s/latest.ckpt' % (args.checkpoint_path))
if one_direction:
G.load_state_dict(ckpt[gen_name])
G.eval()
else:
Gab.load_state_dict(ckpt['Gab'])
Gba.load_state_dict(ckpt['Gba'])
Gab.eval()
Gba.eval()
print('Eval mode')
# +
biopsy_patch_no_map = {}
biopsy_target_map = {}
transform = transforms.Compose([
transforms.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5])
])
for i, patch_name in enumerate(os.listdir(source_path)):
if not folder_to_folder:
# Check if patch should be sent to valid for every patch from new patient
if patch_name.split('__')[0] not in biopsy_patch_no_map:
biopsy_patch_no_map[patch_name.split('__')[0]] = 0
if random.randint(1, 10) > train_valid_split*10:
biopsy_target_map[patch_name.split('__')[0]] = 'train'
else:
biopsy_target_map[patch_name.split('__')[0]] = 'valid'
# Keeping track of number of patches per biopsy crop
biopsy_patch_no_map[patch_name.split('__')[0]] += 1
img = Image.open(source_path + patch_name)
img = img.convert('RGB')
img = img.resize((size, size))
img = np.array(img)
img = img.transpose(2, 0, 1)
img = img / 255.
img = torch.FloatTensor(img).to(device)
image = transform(img)
image = image.unsqueeze(0)
if one_direction or patch_name.startswith('C'):
out = Gba(image)
else:
out = Gab(image)
out = Gba(out)
if not folder_to_folder:
biopsy_target_path = target_path.replace('train', biopsy_target_map[patch_name.split('__')[0]])
torchvision.utils.save_image((out + 1)/2, biopsy_target_path + patch_name)
else:
biopsy_target_path = target_path
torchvision.utils.save_image((out + 1)/2, biopsy_target_path + patch_name)
if i % 1000 == 0:
print(i)
# -
os.listdir(source_path)[0].split('__')