-
Notifications
You must be signed in to change notification settings - Fork 10
/
Evaluation.py
182 lines (158 loc) · 7.97 KB
/
Evaluation.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
import pickle
import itertools
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import roc_curve
from sklearn.metrics import confusion_matrix
import pylab as pl
import csv
import os
class evaluation:
def __init__(self):
return None
def plot_confusion_matrix(self,cm,method,path, classes,normalize=False,title='Confusion matrix',cmap=plt.cm.Blues):
"""
This function prints and plots the confusion matrix.
Normalization can be applied by setting `normalize=True`.
"""
if normalize:
cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
print("Normalized confusion matrix")
else:
print('Confusion matrix, without normalization')
plt.figure()
plt.imshow(cm, interpolation='nearest', cmap=cmap)
plt.title(title)
plt.colorbar()
tick_marks = np.arange(len(classes))
plt.xticks(tick_marks, classes, rotation=45)
plt.yticks(tick_marks, classes)
fmt = '.2f' if normalize else 'd'
thresh = cm.max() / 2.
for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):
plt.text(j, i, format(cm[i, j], fmt),
horizontalalignment="center", fontsize=30,
color="white" if cm[i, j] > thresh else "black")
plt.ylabel('True label',fontsize=20)
plt.xlabel('Predicted label',fontsize=20)
plt.tight_layout()
plt.savefig(os.path.join(path,method +'_ConfMatrix.jpg'))
def get_threshold(self, fprs, thresholds, fpr):
# Getting threshold for particular fpr
threshold = 0
for x in range(0, fprs.size):
if fprs[x] >= fpr:
break
threshold = thresholds[x]
return threshold
def get_result(self, method,imgNames, true_label,predict_score,path, minThreshold = -1):
# Getting predicted scores
predict = np.array(predict_score)
if(len(predict.shape)==2):
predict = predict[:, 1]
elif(len(predict.shape)==3):
predict = predict[:, :, 1]
# Normalization of scores in [0,1]
predictScore = (predict-min(predict))/ (max(predict) - min(predict))
print('Max Score:'+ str(max(predict)))
print('Min Score:'+ str(min(predict)))
# Saving image or video name with match score
if imgNames != 'None':
imgNameScore=[]
for i in range(len(imgNames)):
imgNameScore.append([imgNames[i], true_label[i], predictScore[i]])
with open(os.path.join(path, method + '_Match_Scores.csv'), 'w', newline='') as fout:
writer = csv.writer(fout)
writer.writerows(imgNameScore)
# Histogram plot
live = []
[live.append(predictScore[i]) for i in range(len(true_label)) if (true_label[i] == 0)]
spoof = []
[spoof.append(predictScore[j]) for j in range(len(true_label)) if (true_label[j] == 1)]
bins = np.linspace(np.min(np.array(spoof + live)), np.max(np.array(spoof + live)), 60)
plt.figure()
plt.hist(live, bins, alpha=0.5, label='Bonafide', density=True, edgecolor='black', facecolor='g')
plt.hist(spoof, bins, alpha=0.5, label='PA', density=True, edgecolor='black',facecolor='r' )
plt.legend(loc='upper right', fontsize=15)
plt.xlabel('Scores')
plt.ylabel('Frequency')
plt.savefig(os.path.join(path, method +"_Histogram.jpg"))
# Plot ROC curves in semilog scale
(fprs, tprs, thresholds) = roc_curve(true_label, predictScore)
plt.figure()
plt.semilogx(fprs, tprs, label=method)
plt.grid(True, which="major")
plt.legend(loc='lower right', fontsize=15)
plt.yticks(np.arange(0, 1.1, 0.1))
plt.xticks([0.001, 0.01, 0.1, 1])
plt.xlabel('False Detection Rate')
plt.ylabel('True Detection Rate')
plt.xlim((0.0005, 1.01))
plt.ylim((0, 1.02))
plt.plot([0.002, 0.002], [0, 1], color='#A0A0A0', linestyle='dashed')
plt.plot([0.001, 0.001], [0, 1], color='#A0A0A0', linestyle='dashed')
plt.plot([0.01, 0.01], [0, 1], color='#A0A0A0', linestyle='dashed')
plt.savefig(os.path.join(path,method +"_ROC.jpg"))
#Plot Raw ROC curves
plt.figure()
plt.plot(fprs, tprs)
plt.grid(True, which="major")
plt.legend(method, loc='lower right', fontsize=15)
plt.yticks(np.arange(0, 1.1, 0.1))
plt.xticks([0.01, 0.1, 1])
plt.xlabel('False Detection Rate')
plt.ylabel('True Detection Rate')
plt.xlim((0.0005, 1.01))
plt.ylim((0, 1.02))
plt.savefig(os.path.join(path,method +"_RawROC.jpg"))
# Calculation of TDR at 0.2% , 0.1% and 5% FDR
with open(os.path.join(path , method +'_TDR-ACER.csv'), mode='w+') as fout:
fprArray = [0.002,0.001, 0.01, 0.05]
for fpr in fprArray:
tpr = np.interp(fpr, fprs, tprs)
threshold = self.get_threshold(fprs, thresholds, fpr)
fout.write("TDR @ FDR, threshold: %f @ %f ,%f\n" % (tpr, fpr, threshold))
print("TDR @ FDR, threshold: %f @ %f ,%f " % (tpr, fpr, threshold))
# Calculation of APCER, BPCER and ACER
if minThreshold == -1:
minACER= 1000
for thresh in pl.frange(0,1,0.025):
APCER = np.count_nonzero(np.less(spoof,thresh))/len(spoof)
BPCER = np.count_nonzero(np.greater_equal(live,thresh))/len(live)
ACER = (APCER + BPCER)/2
if ACER < minACER:
minThreshold = thresh
minAPCER = APCER
minBPCER = BPCER
minACER = ACER
fout.write("APCER and BPCER @ ACER, threshold: %f and %f @ %f, %f\n" % (minAPCER, minBPCER, minACER, minThreshold))
print("APCER and BPCER @ ACER, threshold: %f and %f @ %f, %f\n" % (minAPCER, minBPCER, minACER, minThreshold))
else:
APCER = np.count_nonzero(np.less(spoof, minThreshold)) / len(spoof)
BPCER = np.count_nonzero(np.greater_equal(live, minThreshold)) / len(live)
ACER = (APCER + BPCER) / 2
fout.write("APCER and BPCER @ ACER, threshold: %f and %f @ %f, %f\n" % (APCER, BPCER, ACER, minThreshold))
print("APCER and BPCER @ ACER, threshold: %f and %f @ %f, %f\n" % (APCER, BPCER, ACER, minThreshold))
# Calculation of Confusion matrix
#threshold = self.get_threshold(fprs, thresholds, 0.002)
predict = predictScore >= minThreshold
predict_label =[]
[predict_label.append(int(predict[i])) for i in range(len(predict))]
conf_matrix = confusion_matrix(true_label, predict_label) # 0 for live and 1 for spoof
print(conf_matrix)
# Plot non-normalized confusion matrix
np.set_printoptions(precision=2)
class_names = ['0', '1']
self.plot_confusion_matrix(conf_matrix, method,path, classes=class_names,normalize=False)
# Saving evaluation measures
pickle.dump((fprs,tprs,minThreshold,tpr,fpr,conf_matrix), open(os.path.join(path,method +".pickle"), "wb"))
errorIndex=[]
[errorIndex.append(i) for i in range(len(true_label)) if true_label[i] != predict_label[i]]
return errorIndex, predictScore, minThreshold
if __name__ == '__main__':
true_label= [0,0,0,0,0,1,1,1,1,1]
predict_score = [[0.8,0.20], [0.44, 0.56],[0.6,0.4],[0.7, 0.3], [0.9,0.1],[0.2,0.8],[0.3,0.7],[0.4,0.6],[0.01,0.99],[0.1, 0.9]]
obvResult = evaluation()
predict_score = np.array(predict_score)
predict_score = predict_score.reshape((10,1, 2))
result = obvResult.get_result('Test', 'None', true_label, predict_score, "../TempData/results")