-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathmodel.py
191 lines (174 loc) · 8.45 KB
/
model.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
183
184
185
186
187
188
189
190
191
from keras.applications.vgg16 import VGG16
from keras.applications.vgg19 import VGG19
from keras.applications.resnet50 import ResNet50
from keras.applications.densenet import DenseNet121
from keras.applications.xception import Xception
from keras.preprocessing import image
import os, pandas as pd
import numpy as np
from sklearn.cluster import KMeans
class Model:
def __init__(self,model_name):
self.model=model_name
def build_model(self):
if self.model=='Xception' or self.model=='xception':
return Xception(include_top=False,pooling='avg',weights='imagenet'),'Xception'
elif self.model=='DenseNet' or self.model=='densenet':
return DenseNet121(include_top=False,pooling='avg',weights='imagenet'),'DenseNet'
elif self.model=='ResNet' or self.model=='resnet':
return ResNet50(include_top=False,pooling='avg',weights='imagenet'),'ResNet'
elif self.model=='VGG16' or self.model=='vgg16':
return VGG16(include_top=False,pooling='avg',weights='imagenet'),'VGG16'
return VGG19(include_top=False,pooling='avg',weights='imagenet'),'VGG19'
def get_features(img_folder,Model):
fp1=pd.DataFrame(columns=['filename','feature'])
fp2 = pd.DataFrame(columns=['filename', 'feature'])
for i in os.listdir(img_folder):#stage1 stage2
next_folder=os.path.join(img_folder,i)
for j in os.listdir(next_folder):#Belt st1...
next_folder_plus=os.path.join(next_folder,j)
for h in os.listdir(next_folder_plus):
path=os.path.join(next_folder_plus,h)
print(path)
img=image.load_img(path,target_size=(224,224))
x=image.img_to_array(img)
x=np.expand_dims(x,axis=0)
features=Model[0].predict(x)[0]
fp1=fp1.append({'filename':path,'feature':features},ignore_index=True)
img = image.load_img(path)
x = image.img_to_array(img)
x = np.expand_dims(x, axis=0)
features = Model[0].predict(x)[0]
fp2 = fp2.append({'filename': path, 'feature': features}, ignore_index=True)
print(path+' finished!!')
fp1.to_csv('data/{}_features_fixed_size.csv'.format(Model[1]),index=False)
fp2.to_csv('data/{}_features_default_size.csv'.format(Model[1]),index=False)
print('ALL DONE!')
class ImageCluster(object):
def __init__(self,base_img_folder,resorted_img_folder,
cluster_algo='kmeans',base_model='vgg16',k=None,maxK=None,csv_file_path=None):
self.base_model, self.base_model_name = Model(model_name=base_model).build_model()
self.csv_file_path=csv_file_path
self.cluster_algo=cluster_algo
self.k=k
self.maxK=maxK
self.base_img_folder=base_img_folder
self.resorted_img_folder=resorted_img_folder
def get_feature_map(self,resize_shape=None):
"""
You can fix the code as your actual situation and environment!!!
设置图片文件夹一定是3层的，
base_model_folder
----label
----xxxxxxxx.jpg
:param resize_shape:
:return:
"""
img_path_all=[]
f=pd.DataFrame(columns=['filename','feature'])
if resize_shape==None:
if os.path.isdir(self.base_img_folder):
for i in os.listdir(self.base_img_folder):
next_path=os.path.join(os.path.join(self.base_img_folder,i))
if os.path.isdir(next_path):
for j in os.listdir(next_path):
last_path = os.path.join(next_path, j)
for w in os.listdir(last_path):
img_path = os.path.join(last_path, w)
img_path_all.append(img_path)
img = image.load_img(img_path, target_size=resize_shape)
x = image.img_to_array(img)
x = np.expand_dims(x, axis=0)
features = self.base_model.predict(x)[0]
f = f.append({'filename': img_path, 'feature': features}, ignore_index=True)
else:
pass
else:
raise ValueError('the base image folder is wrong! PLZ check it out')
else:
if os.path.isdir(self.base_img_folder):
for i in os.listdir(self.base_img_folder):
next_path=os.path.join(os.path.join(self.base_img_folder,i))
if os.path.isdir(next_path):
for j in os.listdir(next_path):
last_path = os.path.join(next_path, j)
for w in os.listdir(last_path):
img_path=os.path.join(last_path,w)
img_path_all.append(img_path)
img = image.load_img(img_path, target_size=resize_shape)
x = image.img_to_array(img)
x = np.expand_dims(x, axis=0)
features = self.base_model.predict(x)[0]
f = f.append({'filename': img_path, 'feature': features}, ignore_index=True)
print(img_path,' extracted features')
else:
pass
else:
raise ValueError('the base image folder is wrong! PLZ check it out')
if len(img_path_all)==0:
raise ValueError('image loading fails,please check you image path!')
else:
print('Have got the feature map for each image')
f.to_csv('output/base_model_{}_feature_maps.csv'.format(self.base_model_name))
print('output/base_model_{}_feature_maps.csv has finished!'.format(self.base_model_name))
def kmeans(self):
x=[]
if self.csv_file_path == None:
self.csv_file = pd.read_csv('output/base_model_{}_feature_maps.csv'.format(self.base_model_name))
else:
self.csv_file = pd.read_csv(self.csv_file_path)
for i in self.csv_file['feature']:
x.append([float(t) for t in i.strip('[').strip(']').split(' ')])
x=np.array(x)
if os.path.exists('output'):
pass
else:
os.mkdir('output')
if os.path.exists('matplot'):
pass
else:
os.mkdir('matplot')
def func(k):
model = KMeans(n_clusters=k, init='k-means++')
model.fit(x)
# print('cluster_center', model.cluster_centers_)
f = pd.DataFrame(columns=['filename', 'label'])
f['filename'] = self.csv_file['filename']
f['label'] = model.labels_
f.to_csv('output/base_model_{}_cluster_kmeans_{}.csv'.format(self.base_model_name,str(k)))
return model.inertia_
if self.k==None:
sse=[]
for k in range(51,self.maxK+1):
sse.append(func(k))
import matplotlib.pyplot as plt
plt.plot(range(2,self.maxK+1),sse,marker='o')
plt.xlabel('number of K(cluster)')
plt.ylabel('SSE Value for each K')
plt.title('KMeans for ImageCluster')
plt.savefig('matplot/base_model_{}_KMeans_maxK_{}.png'.format(self.base_model_name,str(self.maxK)))
plt.show()
else:
func(self.k)
def imagecluster(self):
if self.cluster_algo.lower()=='kmeans':
self.kmeans()
else:
print('no existing cluster algorithm')
return
def resorted_img(self,selected_k_num):
import shutil
if os.path.exists(self.resorted_img_folder):
pass
else:
os.mkdir(self.resorted_img_folder)
resorted_csv=pd.read_csv('output/base_model_{}_cluster_kmeans_{}.csv'.format(self.base_model_name,str(selected_k_num)))
for i in resorted_csv.index:
filename=resorted_csv.loc[i,'filename']
label=resorted_csv.loc[i,'label']
if os.path.exists(os.path.join(self.resorted_img_folder,str(label))):
pass
else:
os.mkdir(os.path.join(self.resorted_img_folder,str(label)))
shutil.copy(filename,os.path.join(self.resorted_img_folder,str(label)))
print(os.path.join(self.resorted_img_folder,str(label))+'\\'+filename+' Copied！')