/
session_cluster.py
142 lines (113 loc) · 4.89 KB
/
session_cluster.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
import cv2
import numpy as np
import cPickle
import time
from sklearn.preprocessing import StandardScaler
from sklearn import svm
from sklearn.decomposition import PCA
import sys
def get_dataset():
train_images_filenames = cPickle.load(open('train_images_filenames.dat','r'))
train_images_filenames = [filename.replace('../../Databases/', '') for filename in train_images_filenames]
test_images_filenames = cPickle.load(open('test_images_filenames.dat','r'))
test_images_filenames = [filename.replace('../../Databases/', '') for filename in test_images_filenames]
train_labels = cPickle.load(open('train_labels.dat','r'))
test_labels = cPickle.load(open('test_labels.dat','r'))
print 'Loaded '+str(len(train_images_filenames))+' training images filenames with classes ',set(train_labels)
print 'Loaded '+str(len(test_images_filenames))+' testing images filenames with classes ',set(test_labels)
return train_images_filenames, test_images_filenames, train_labels, test_labels
def get_feature_detector(name='sift', n_features=100):
if name == 'sift':
return cv2.SIFT(nfeatures=n_features)
elif name == 'orb':
return cv2.ORB(nfeatures=n_features)
elif name == 'surf':
return cv2.SURF(hessianThreshold=400)
else:
raise NotImplemented
def extract_features(FEATdetector, train_images_filenames, train_labels, nImages):
Train_descriptors = []
Train_label_per_descriptor = []
for i in range(len(train_images_filenames)):
filename=train_images_filenames[i]
if Train_label_per_descriptor.count(train_labels[i])<nImages:
print 'Reading image '+filename
ima=cv2.imread(filename)
gray=cv2.cvtColor(ima,cv2.COLOR_BGR2GRAY)
kpt,des=FEATdetector.detectAndCompute(gray,None)
Train_descriptors.append(des)
Train_label_per_descriptor.append(train_labels[i])
print str(len(kpt))+' extracted keypoints and descriptors'
# Transform everything to numpy arrays
D=Train_descriptors[0]
L=np.array([Train_label_per_descriptor[0]]*Train_descriptors[0].shape[0])
for i in range(1,len(Train_descriptors)):
D=np.vstack((D,Train_descriptors[i]))
L=np.hstack((L,np.array([Train_label_per_descriptor[i]]*Train_descriptors[i].shape[0])))
return D, L
def train_SVM(kernel, C, D, L):
stdSlr = StandardScaler().fit(D)
D_scaled = stdSlr.transform(D)
print 'Training the SVM classifier...'
clf = svm.SVC(kernel='linear', C=1).fit(D_scaled, L)
print 'Done!'
return clf, stdSlr
def test_SVM(FEATdetector, test_images_filenames, test_labels, clf, stdSlr, reducer):
numtestimages=0
numcorrect=0
for i in range(len(test_images_filenames)):
filename=test_images_filenames[i]
ima=cv2.imread(filename)
gray=cv2.cvtColor(ima,cv2.COLOR_BGR2GRAY)
kpt,des=FEATdetector.detectAndCompute(gray,None)
if reducer:
des = reducer.transform(des)
predictions = clf.predict(stdSlr.transform(des))
values, pos = np.unique(predictions, return_inverse=True)
counts = np.bincount(pos)
predictedclass = values[np.argmax(counts)]
print 'image '+filename+' was from class '+test_labels[i]+' and was predicted '+predictedclass
numtestimages+=1
if predictedclass==test_labels[i]:
numcorrect+=1
return numcorrect, numtestimages
def PCA_reduce(D, n_components):
print(D.shape)
pca = PCA(n_components=n_components)
pca.fit(D)
return pca.transform(D), pca
def main(nfeatures=100, nImages=30, n_components=20, kernel='linear', C=1, reduction=None, features='sift', outFile='out'):
start = time.time()
# read the train and test files
train_images_filenames, test_images_filenames, train_labels, test_labels = get_dataset()
# create the SIFT detector object
FEATdetector = get_feature_detector(name=features, n_features=nfeatures)
# read the just 30 train images per class
# extract SIFT keypoints and descriptors
# store descriptors in a python list of numpy arrays
D, L = extract_features(FEATdetector, train_images_filenames, train_labels, nImages)
if reduction == 'pca':
D, reducer = PCA_reduce(D, n_components)
else:
reducer = None
print(D.shape)
# Train a linear SVM classifier
clf, stdSlr = train_SVM(kernel, C, D, L)
# get all the test data and predict their labels
numcorrect, numtestimages = test_SVM(FEATdetector, test_images_filenames, test_labels, clf, stdSlr, reducer)
print 'Final accuracy: ' + str(numcorrect*100.0/numtestimages)
end=time.time()
print 'Done in '+str(end-start)+' secs.'
f = open(outFile,'a')
f.write('Config= nfeatures:'+nfeatures+' nImages:'+nImages+' n_components:'+n_components+' kernel:'+kernel+ ' c:'+c+ ' reduction:'+reduction+' features:'+features+ ' Final accuracy= ' + str(numcorrect*100.0/numtestimages) + ' Done in '+str(end-start)+' secs.' )
f.close()
## 38.78% in 797 secs.
nfeatures = int(sys.argv[1])
nImages = int(sys.argv[2])
n_components = int(sys.argv[3])
kernel = str(sys.argv[4])
c = int(sys.argv[5])
reduction = str(sys.argv[6])
features = str(sys.argv[7])
outFile = str(sys.argv[8])
main(nfeatures, nImages, n_components, kernel, c, reduction, features, outFile)