1 import tensorflow as tf
 2 import os
 3 import numpy as np
 4 import re
 5 from PIL import Image
 6 import matplotlib.pyplot as plt
 7
 8 class NodeLookup(object):
 9     def __init__(self):
10         label_lookup_path = 'inception_model/imagenet_2012_challenge_label_map_proto.pbtxt'
11         uid_lookup_path = 'inception_model/imagenet_synset_to_human_label_map.txt'
12         self.node_lookup = self.load(label_lookup_path, uid_lookup_path)
13
14     def load(self, label_lookup_path, uid_lookup_path):
15         # 加载分类字符串n********对应分类名称的文件
16         proto_as_ascii_lines = tf.gfile.GFile(uid_lookup_path).readlines()
17         uid_to_human = {}
18         #一行一行读取数据
19         for line in proto_as_ascii_lines :
20             #去掉换行符
21             line=line.strip('\n')
22             #按照'\t'分割
23             parsed_items = line.split('\t')
24             #获取分类编号
25             uid = parsed_items[0]
26             #获取分类名称
27             human_string = parsed_items[1]
28             #保存编号字符串n********与分类名称映射关系
29             uid_to_human[uid] = human_string
30
31         # 加载分类字符串n********对应分类编号1-1000的文件
32         proto_as_ascii = tf.gfile.GFile(label_lookup_path).readlines()
33         node_id_to_uid = {}
34         for line in proto_as_ascii:
35             if line.startswith('  target_class:'):
36                 #获取分类编号1-1000
37                 target_class = int(line.split(': ')[1])
38             if line.startswith('  target_class_string:'):
39                 #获取编号字符串n********
40                 target_class_string = line.split(': ')[1]
41                 #保存分类编号1-1000与编号字符串n********映射关系
42                 node_id_to_uid[target_class] = target_class_string[1:-2]  #要将两侧的双引号去掉
43
44         #建立分类编号1-1000对应分类名称的映射关系
45         node_id_to_name = {}
46         for key, val in node_id_to_uid.items():
47             #获取分类名称
48             name = uid_to_human[val]
49             #建立分类编号1-1000到分类名称的映射关系
50             node_id_to_name[key] = name
51         return node_id_to_name
52
53     #传入分类编号1-1000返回分类名称
54     def id_to_string(self, node_id):
55         if node_id not in self.node_lookup:
56             return ''
57         return self.node_lookup[node_id]
58
59
60 #创建一个图来存放google训练好的模型
61 with tf.gfile.FastGFile('inception_model/classify_image_graph_def.pb', 'rb') as f:
62     graph_def = tf.GraphDef()
63     graph_def.ParseFromString(f.read())
64     tf.import_graph_def(graph_def, name='')
65
66
67 with tf.Session() as sess:
68     softmax_tensor = sess.graph.get_tensor_by_name('softmax:0')
69     #遍历目录
70     for root,dirs,files in os.walk('images/'):
71         for file in files:
72             #载入图片
73             image_data = tf.gfile.FastGFile(os.path.join(root,file), 'rb').read()
74             predictions = sess.run(softmax_tensor,{'DecodeJpeg/contents:0': image_data})#图片格式是jpg格式
75             predictions = np.squeeze(predictions)#把结果转为1维数据
76
77             #打印图片路径及名称
78             image_path = os.path.join(root,file)
79             print(image_path)
80             #显示图片
81             img=Image.open(image_path)
82             plt.imshow(img)
83             plt.axis('off')
84             plt.show()
85
86             #排序
87             top_k = predictions.argsort()[-5:][::-1]
88             node_lookup = NodeLookup()
89             for node_id in top_k:
90                 #获取分类名称
91                 human_string = node_lookup.id_to_string(node_id)
92                 #获取该分类的置信度
93                 score = predictions[node_id]
94                 print('%s (score = %.5f)' % (human_string, score))
95             print()

        proto_as_ascii_lines = tf.gfile.GFile(uid_lookup_path).readlines()   uid_to_human = {}     #一行一行读取数据for line in proto_as_ascii_lines :#去掉换行符line=line.strip('\n')#按照'\t'分割parsed_items = line.split('\t')#获取分类编号uid = parsed_items[0]#获取分类名称human_string = parsed_items[1]#保存编号字符串n********与分类名称映射关系uid_to_human[uid] = human_string

        # 加载分类字符串n********对应分类编号1-1000的文件proto_as_ascii = tf.gfile.GFile(label_lookup_path).readlines()node_id_to_uid = {}for line in proto_as_ascii:if line.startswith('  target_class:'):#获取分类编号1-1000target_class = int(line.split(': ')[1])if line.startswith('  target_class_string:'):#获取编号字符串n********target_class_string = line.split(': ')[1]#保存分类编号1-1000与编号字符串n********映射关系node_id_to_uid[target_class] = target_class_string[1:-2]