-
Notifications
You must be signed in to change notification settings - Fork 0
/
HNSW_query.py
97 lines (85 loc) · 3.26 KB
/
HNSW_query.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
#
# ___====-_ _-====___
# _--^^^#####// \\#####^^^--_
# _-^##########// ( ) \\##########^-_
# -############// |\^^/| \\############-
# _/############// (@::@) \\############\_
# /#############(( \\// ))#############\
# -###############\\ (oo) //###############-
# -#################\\ / VV \ //#################-
# -###################\\/ \//###################-
# _#/|##########/\######( /\ )######/\##########|\#_
# |/ |#/\#/\#/\/ \#/\##\ | | /##/\#/ \/\#/\#/\#| \|
# ` |/ V V ` V \#\| | | |/#/ V ' V V \| '
# ` ` ` ` / | | | | \ ' ' ' '
# ( | | | | )
# __\ | | | | /__
# (vvv(VVV)(VVV)vvv)
#
# God bless me, no bug!
# `=---='
#
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
import time
import h5py
from PIL import Image
import hnswlib
import numpy as np
import os
import argparse
from extract_cnn_vgg16_keras import VGGNet
os.environ["CUDA_VISIBLE_DEVICES"]="2"
# read in indexed images' feature vectors and corresponding image names加载数据
ap = argparse.ArgumentParser()
ap.add_argument("-query", required=True,
help="Path to query which contains image to be queried")
args = vars(ap.parse_args())
h5f = h5py.File('feature0416.h5', 'r')
imgNames = h5f['dataset_2'][:]
h5f.close()
num_elements = len(imgNames)
print('\n检索图片库的总数：{}张\n'.format(num_elements))
labels_index = np.arange(num_elements)
#EMBEDDING_SIZE = feats.shape[1]
p = hnswlib.Index(space = 'cosine', dim = 512) # possible options are l2, cosine or ip
print('&&&&&&&&&&&&&&&&&&&')
time0 = time.time()
p.load_index('index_one_million.idx')
time01 = time.time()
print('\n加载模型所用时间：{}秒\n'.format(time01-time0))
print("*******************")
p.set_ef(300) # ef should always be > k
model = VGGNet()
time_vgg2 = time.time()
print('\n启动VGG_NET用时：{}秒\n'.format(time_vgg2-time01))
# from extract_cnn_vgg16_keras import VGGNet
queryDir = args["query"]
# queryImg = mpimg.imread(queryDir) # 读取图片
# plt.title("Query Image")
# plt.imshow(queryImg)
# plt.show()
# init VGGNet16 model
time2 = time.time()
# extract query image's feature, compute simlarity score and sort
queryVec = model.extract_feat(queryDir)
# time02 = time.time()
# print('\n启动配置环境 + 提取特征所用总时间：{}秒\n'.format(time02-time2))
time3 = time.time()
labels, distances = p.knn_query(queryVec, k = 100)
# print(distances)
imlist = [imgNames[index] for i, index in enumerate(labels)][0]
time03 = time.time()
print('\n检索100张相似图所用时间：{}秒\n'.format(time03-time3))
print('\n总用时间：{}秒\n'.format(time03-time2))
for i, im in enumerate(imlist):
im=im.decode('utf-8')
# image = mpimg.imread(im)
img = Image.open(im)
img.save('./1/%s.jpg' % i)
# plt.title("search output %d" % (i + 1))
# plt.imshow(image)
# plt.show()
print('\n*****完成*****\n')
# # images = [queryImg]
# # images += [plt.imread(imgNames for label in labels[0])]
# plot_predictions(images)