-
Notifications
You must be signed in to change notification settings - Fork 2
/
analyse_featmap_adv.py
132 lines (108 loc) · 5.18 KB
/
analyse_featmap_adv.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
"""Detect adv/clean from the hidden feature"""
from __future__ import absolute_import
from __future__ import print_function
import os
import argparse
from datasets import get_data
from models import get_model
import numpy as np
import sklearn.metrics
import keras.backend as K
import keras
from sklearn.model_selection import StratifiedShuffleSplit, StratifiedKFold, train_test_split
from sklearn.decomposition import PCA
from sklearn.svm import SVC
from sklearn.metrics import accuracy_score, roc_auc_score, roc_curve
from sklearn.ensemble import RandomForestClassifier
import matplotlib.pyplot as plt
from global_config import *
import glob
import imageio
import cv2
from vis.visualization import visualize_saliency
from vis.utils import utils
DATASETS = ['dr', 'cxr', 'derm']
ATTACKS = ['fgsm', 'bim', 'jsma', 'cw-l2', 'clean']
TEST_SIZE = {'dr': 0.7, 'cxr': 0.7, 'derm': 0.5}
CLIP_MIN = {'mnist': -0.5, 'cifar': -0.5, 'svhn': -0.5, 'dr': -1.0, 'cxr': -1.0, 'derm': -1.0, 'imagenet':-128}
CLIP_MAX = {'mnist': 0.5, 'cifar': 0.5, 'svhn': 0.5, 'dr': 1.0, 'cxr': 1.0, 'derm': 1.0, 'imagenet':128}
def solve_name_controdiction(model):
for name in map(lambda x: x.__class__.__name__, model.layers):
K.get_uid(name)
K.get_uid('input')
K.get_uid('input')
def analyze(args):
assert args.dataset in ['mnist', 'cifar-10', 'svhn', 'dr', 'cxr', 'derm', 'imagenet'], \
"Dataset parameter must be either 'mnist', 'cifar-10', 'svhn', 'dr', 'cxr', or 'derm'"
# load feature/label data
if args.dataset == 'imagenet':
flist = sorted(glob.glob('data/imagenet/*.jpg'))
cX = map(lambda path: cv2.resize(imageio.imread(path), (224, 224)), flist)
cX = list(cX)
cX = np.stack(cX)
cX = keras.applications.resnet50.preprocess_input(cX)
aX = np.load('data/eps1_Adv_imagenet_pgd.npy')
model = keras.applications.resnet50.ResNet50(include_top=True)
cy = model.predict(cX).argmax(-1) #[248, 281, 281, 248, 281]
ay = model.predict(aX).argmax(-1)
elif args.dataset == 'mnist':
_, _, X, y = get_data(args.dataset, onehot=False, split_traintest=False) # clean image
model = get_model(args.dataset, softmax=False)
layer_idx = utils.find_layer_idx(model, 'dense_2')
solve_name_controdiction(model)
else:
_, _, cX, cy = get_data(args.dataset, onehot=False, split_traintest=False) # clean image
aX = np.load('data/eps1_Adv_%s_pgd.npy' % args.dataset)
model = get_model(args.dataset)
solve_name_controdiction(model)
ay = model.predict(aX, verbose=1).argmax(-1)
x_in = model.input
resfeatmap = model.get_layer('res5c_branch2a').input
featmodel = keras.Model(x_in, resfeatmap)
def reg(x, min=None, max=None):
if x.shape[-1] == 1:
x = np.tile(x, [1, 1, 3])
max = max or x.max()
min = min or x.min()
return np.clip((x - min) / (max - min), 0., 1.)
plot_range = {
'imagenet': slice(0, None), # 2
'derm': np.concatenate([ np.where(cy>0)[0][-5:], np.where(cy<1)[0][:5] ]), # -1
'dr': np.concatenate([ np.where(cy>0)[0][-5:], np.where(cy<1)[0][:5] ]), # -6
'cxr': np.concatenate([ np.where(cy>0)[0][-5:], np.where(cy<1)[0][:5] ]), #-
'mnist': np.where(cy == 6)[0][:10],
}
def do_analyze(X, y, advclean='clean'):
for i, (img, label) in enumerate(zip(X[plot_range[args.dataset]], y[plot_range[args.dataset]])):
featmap, = featmodel.predict(img[None, ...])
if args.dataset == 'imagenet':
img = img[..., ::-1]
print('min, max = ', np.min(np.mean(featmap, axis=-1)), np.max(np.mean(featmap, axis=-1)))
featmap_c = cv2.applyColorMap(((1 - reg(np.mean(featmap, axis=-1), 0.2, 1.5)) * 255.9).astype('uint8'), cv2.COLORMAP_JET)
featmap_c = cv2.resize(featmap_c, (512, 512), interpolation=cv2.INTER_NEAREST)
imageio.imwrite('vis/featmap_adv/%s_%d_%s_y=%d_featmap.png' % (args.dataset, i, advclean, label), featmap_c)
imageio.imwrite('vis/featmap_adv/%s_%d_%s_y=%d_original.png' % (args.dataset, i, advclean, label), reg(img))
# plt.imshow(np.mean(featmap, axis=-1), cmap='jet');plt.colorbar();plt.axis('off');plt.savefig('vis/featmap_adv/%s_%d_%s_y=%d_featmap.png' % (args.dataset, i, advclean, label));plt.show()
# plt.imshow(reg(img));imageio.imwrite('vis/featmap_adv/%s_%d_%s_y=%d_original.png' % (args.dataset, i, advclean, label), reg(img));plt.show()
do_analyze(cX, cy, 'clean')
do_analyze(aX, ay, 'adv')
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument(
'-d', '--dataset',
help="Dataset to use",
required=True, type=str
)
parser.add_argument(
'-a', '--attack',
help="Attack to use train the discriminator; either 'fgsm', 'bim-a', 'bim-b', 'jsma', 'cw-l2'",
required=False, type=str
)
# args = parser.parse_args()
# analyze(args)
for ds in ['imagenet', 'dr', 'cxr', 'derm']:
# for atk in ['fgsm', 'bim', 'pgd']:
argv = ['-d', ds]
print('\n$> ', argv)
args = parser.parse_args(argv)
analyze(args)