python3下tensorflow练习（六）之VGG16/VGG19探索

VGG16/19是ImageNet比赛中获胜的经典之作，学习该算法的思想来入门深度学习很有必要。而且当我们自己的训练数据量过小时，往往借助牛人已经预训练好的网络进行特征的提取，然后在后面加上自己特定任务的网络进行调优。

网络结构如下：

VGG16网上讲的很详细了，我这就不过多解释了，这次主要讲解一下VGG16的源码，以及怎么利用github的开源项目，开启自己的项目。

小伙伴可以通过以下链接下载源码：VGG16/19源码

网络结构vgg16.py代码，如下：

import inspect###获取加载模块路径
import os

import numpy as np
import tensorflow as tf
import time

VGG_MEAN = [103.939, 116.779, 123.68]

"""   
numpy将数组以二进制格式保存到磁盘
np.load和np.save是读写磁盘数组数据的两个主要函数，默认情况下，数组是以未压缩的原始二进制格式保存在扩展名为.npy的文件中。
np.save("A.npy",A)   #如果文件路径末尾没有扩展名.npy，该扩展名会被自动加上。
 B=np.load("A.npy")        
item()方法把字典中每对key和value组成一个元组，并把这些元组放在列表中返回。
person={'name':'lizhong','age':'26','city':'BeiJing','blog':'www.jb51.net'}    
for x in person.items():
    print x
显示：[（'name'，'lizhong'） etc ]
"""
class Vgg16:
    def __init__(self, vgg16_npy_path=None):##构造函数
        if vgg16_npy_path is None:
            path = inspect.getfile(Vgg16)#获得Vgg16--该类的路径
            path = os.path.abspath(os.path.join(path, os.pardir))#os.path.abspath(path) #返回绝对路径，os.pardir为当前目录的父目录
            path = os.path.join(path, "vgg16.npy")#"vgg16.npy"文件名和路径合并成新路径
            vgg16_npy_path = path
            print(path)#打印路径

        self.data_dict = np.load(vgg16_npy_path, encoding='latin1').item()
      #  print (self.data_dict.type())###查看数据类型
        print("npy file loaded")

    def build(self, rgb):
        """
        load variable from npy to build the VGG

        :param rgb: rgb image [batch, height, width, 3] values scaled [0, 1]
        """
  
        start_time = time.time()
        print("build model started")
        rgb_scaled = rgb * 255.0

        # Convert RGB to BGR
        red, green, blue = tf.split(axis=3, num_or_size_splits=3, value=rgb_scaled)
        assert red.get_shape().as_list()[1:] == [224, 224, 1]
        assert green.get_shape().as_list()[1:] == [224, 224, 1]
        assert blue.get_shape().as_list()[1:] == [224, 224, 1]
        bgr = tf.concat(axis=3, values=[
            blue - VGG_MEAN[0],
            green - VGG_MEAN[1],
            red - VGG_MEAN[2],
        ])
        assert bgr.get_shape().as_list()[1:] == [224, 224, 3]

        self.conv1_1 = self.conv_layer(bgr, "conv1_1")#this->conv1_1------新的变量，this->conv_layer---调用成员函数
        self.conv1_2 = self.conv_layer(self.conv1_1, "conv1_2")############self.conv1_2相当于一个成员变量
        self.pool1 = self.max_pool(self.conv1_2, 'pool1')

        self.conv2_1 = self.conv_layer(self.pool1, "conv2_1")
        self.conv2_2 = self.conv_layer(self.conv2_1, "conv2_2")
        self.pool2 = self.max_pool(self.conv2_2, 'pool2')

        self.conv3_1 = self.conv_layer(self.pool2, "conv3_1")
        self.conv3_2 = self.conv_layer(self.conv3_1, "conv3_2")
        self.conv3_3 = self.conv_layer(self.conv3_2, "conv3_3")
        self.pool3 = self.max_pool(self.conv3_3, 'pool3')

        self.conv4_1 = self.conv_layer(self.pool3, "conv4_1")
        self.conv4_2 = self.conv_layer(self.conv4_1, "conv4_2")
        self.conv4_3 = self.conv_layer(self.conv4_2, "conv4_3")
        self.pool4 = self.max_pool(self.conv4_3, 'pool4')

        self.conv5_1 = self.conv_layer(self.pool4, "conv5_1")    
        self.conv5_2 = self.conv_layer(self.conv5_1, "conv5_2")
        self.conv5_3 = self.conv_layer(self.conv5_2, "conv5_3")
        self.pool5 = self.max_pool(self.conv5_3, 'pool5')

        self.fc6 = self.fc_layer(self.pool5, "fc6")
        assert self.fc6.get_shape().as_list()[1:] == [4096]
        self.relu6 = tf.nn.relu(self.fc6)

        self.fc7 = self.fc_layer(self.relu6, "fc7")
        self.relu7 = tf.nn.relu(self.fc7)

        self.fc8 = self.fc_layer(self.relu7, "fc8")

        self.prob = tf.nn.softmax(self.fc8, name="prob")###预测,prob------全局变量
        self.data_dict = None
        print(("build model finished: %ds" % (time.time() - start_time)))

    def avg_pool(self, bottom, name):######bottom----rgb,name------"xxx"
        return tf.nn.avg_pool(bottom, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME', name=name)

    def max_pool(self, bottom, name):
        return tf.nn.max_pool(bottom, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME', name=name)
######padding='SAME'只是保留边缘信息，还是缩减采样
    def conv_layer(self, bottom, name):######bottom----rgb,name------"xxx"
        with tf.variable_scope(name):
            filt = self.get_conv_filter(name)#######得到权重W(四维张量)---vgg16.npy(通过"conv1_1"[0]取出W，也就是filter)

            conv = tf.nn.conv2d(bottom, filt, [1, 1, 1, 1], padding='SAME')

            conv_biases = self.get_bias(name)#######得到偏差b--vgg16.npy(通过"conv1_1"[1]取出b，也就是bias)
            bias = tf.nn.bias_add(conv, conv_biases)

            relu = tf.nn.relu(bias)
            return relu

    def fc_layer(self, bottom, name):
        with tf.variable_scope(name):
            shape = bottom.get_shape().as_list()
            dim = 1
            for d in shape[1:]:
                dim *= d
            x = tf.reshape(bottom, [-1, dim])

            weights = self.get_fc_weight(name)
            biases = self.get_bias(name)

            # Fully connected layer. Note that the '+' operation automatically
            # broadcasts the biases.
            fc = tf.nn.bias_add(tf.matmul(x, weights), biases)

            return fc

    def get_conv_filter(self, name):#######得到权重W(四维张量)---vgg16.npy(通过"conv1_1"取出W，也就是filter)
        return tf.constant(self.data_dict[name][0], name="filter")

    def get_bias(self, name):#######得到偏差b--vgg16.npy(通过"conv1_1"[1]取出b，也就是bias)
        return tf.constant(self.data_dict[name][1], name="biases")

    def get_fc_weight(self, name):
        return tf.constant(self.data_dict[name][0], name="weights")

该网络的特点：

1）用了13个卷积层以及3个全连接层构成VGG16

2)卷积层全是：CONV=3*3 filter, s=1, padding="same"，保持图片形状不变

   每次卷积完了以后用MAX_POOL进行下采样，减小图片尺寸，加快计算速度

   MAX_POOL=2*2 ,s=2

注：该文件自定义类class Vgg()的构造函数是为了获取之前训练好的参数包括权重-----W,和偏差-----bias。这是1000个种类识别的参数比较大，想测试的小伙伴可以通过以下链接下载vgg16.npy文件：下载链接

看一下测试文件test_vgg16.py:

import numpy as np
import tensorflow as tf

#import vgg16
from vgg16 import Vgg16
import utils
import hello

hello.hello()

img1 = utils.load_image("./test_data/tiger.jpeg")
img2 = utils.load_image("./test_data/timg.jpg")

batch1 = img1.reshape((1, 224, 224, 3))
batch2 = img2.reshape((1, 224, 224, 3))

batch = np.concatenate((batch1, batch2), 0)######[2, 224, 224, 3]
'''
numpy.concatenate((a1, a2, ...), axis=0)
0-------列向量
a1, a2, ... : 需要拼接的矩阵
axis : 沿着某个轴拼接，默认为列方向
'''
print (batch.shape)

#with tf.Session(config=tf.ConfigProto(gpu_options=(tf.GPUOptions(per_process_gpu_memory_fraction=0.7)))) as sess:
with tf.device('/cpu:0'):
    with tf.Session() as sess:
        images = tf.placeholder("float", [2, 224, 224, 3])
        feed_dict = {images: batch}

        vgg = Vgg16()##引用vgg16.py里面定义的Vgg16类
        with tf.name_scope("content_vgg"):
            vgg.build(images)

        prob = sess.run(vgg.prob, feed_dict=feed_dict)### 在build()函数里面，属于成员变量，self.prob = tf.nn.softmax(self.fc8, name="prob")
        print(prob)####打印的预测矩阵（2,1000）类似手写字识别： [1,0,0,0...]:表示0
        utils.print_prob(prob[0], './synset.txt')
        utils.print_prob(prob[1], './synset.txt')

测试两张图片：

运行test_vgg16.py文件：

可以看出预测出来了准确的类别：tiger(老虎)和brambling（燕雀）