Tensorflow训练识别手写数字0-9

1.安装环境

这个比较简单，

1.1 安装cnetos7 这个版本中直接代有python2.7.5版本，（下载ISO安装包安装即可我用的是vmware12.5）

1.2 安装 tensorflow

安装pip

yum update -y && yum install -y python python-devel epel-release.noarch python-pip

使用pip安装tensorflow

pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.5.0-cp27-none-linux_x86_64.whl
1.3 安装 python flaskapi

pip install flask(这个不记得了，不行就度娘吧)

1.5 下载MNIST训练库

mnist库
https://files.cnblogs.com/files/keim/train-images-idx3-ubyte.gz.rar 这个文件后缀Rar去掉
https://files.cnblogs.com/files/keim/MNIST_data1.rar 解压和上面的放一起即可

2.训练代码

如下是训练代码，其中mnist_data为上面的MNIST库的位置

#coding=utf-8
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)

import tensorflow as tf

sess = tf.InteractiveSession()

x = tf.placeholder(tf.float32, shape=[None, 784])
y_ = tf.placeholder(tf.float32, shape=[None, 10])
W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))

sess.run(tf.global_variables_initializer())

y = tf.matmul(x,W) + b

cross_entropy = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y))

train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)

for _ in range(1000):
batch = mnist.train.next_batch(100)
train_step.run(feed_dict={x: batch[0], y_: batch[1]})
correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(y_,1))

accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

print(accuracy.eval(feed_dict={x: mnist.test.images, y_: mnist.test.labels}))

def weight_variable(shape):
initial = tf.truncated_normal(shape, stddev=0.1)
return tf.Variable(initial)

def bias_variable(shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)

def conv2d(x, W):
return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')

def max_pool_2x2(x):
return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1], padding='SAME')

W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])
x_image = tf.reshape(x, [-1,28,28,1])
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1)

W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])
h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2)

W_fc1 = weight_variable([7 * 7 * 64, 1024])
b_fc1 = bias_variable([1024])

h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)
keep_prob = tf.placeholder(tf.float32)
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)
W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])

y_conv = tf.matmul(h_fc1_drop, W_fc2) + b_fc2
cross_entropy = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y_conv))
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

saver = tf.train.Saver()  # defaults to saving all variables

sess.run(tf.global_variables_initializer())
for i in range(5000000):
batch = mnist.train.next_batch(50)
if i%100 == 0:
train_accuracy = accuracy.eval(feed_dict={
x:batch[0], y_: batch[1], keep_prob: 1.0})
print("step %d, training accuracy %g"%(i, train_accuracy))

train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})

writer=tf.summary.FileWriter("Scripts",tf.get_default_graph())
writer.close()
print ('save file')
saver.save(sess, 'learning_tensorflow/model.ckpt')  #保存模型参数，注意把这里改为自己的路径
print ('save file ok')
#print("test accuracy %g"%accuracy.eval(feed_dict={x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0}))

3.测试代码

#coding=utf-8
from PIL import Image, ImageFilter
import tensorflow as tf
#import matplotlib.pyplot as plt
import cv2

def imageprepare():
"""
This function returns the pixel values.
The imput is a png file location.
"""
file_name='pic_data/3.png'#导入自己的图片地址
#in terminal 'mogrify -format png *.jpg' convert jpg to png
im = Image.open(file_name).convert('L')
#im.save("pic_data/sample.png")
#plt.imshow(im)
#plt.show()
tv = list(im.getdata()) #get pixel values

#normalize pixels to 0 and 1. 0 is pure white, 1 is pure black.
tva = [ (255-x)*1.0/255.0 for x in tv]
#print(tva)
return tva

# Define the model (same as when creating the model file)
result=imageprepare()
x = tf.placeholder(tf.float32, [None, 784])
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))

def weight_variable(shape):
initial = tf.truncated_normal(shape, stddev=0.1)
return tf.Variable(initial)

def bias_variable(shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)

def conv2d(x, W):
return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')

def max_pool_2x2(x):
return tf.nn.max_pool(x, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')

W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])

x_image = tf.reshape(x, [-1,28,28,1])
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1)

W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])

h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2)

W_fc1 = weight_variable([7 * 7 * 64, 1024])
b_fc1 = bias_variable([1024])

h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

keep_prob = tf.placeholder(tf.float32)
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])

y_conv=tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)

init_op = tf.global_variables_initializer()

saver = tf.train.Saver()
with tf.Session() as sess:
sess.run(init_op)
saver.restore(sess, "learning_tensorflow/model.ckpt")#这里使用了之前保存的模型参数
#print ("Model restored.")

prediction=tf.argmax(y_conv,1)
predint=prediction.eval(feed_dict={x: [result],keep_prob: 1.0}, session=sess)
print(h_conv2)

print('recognize result:')
print(predint[0])

4.结合API远程调用

接口代码：

# coding=UTF-8
from flask import Flask,jsonify,request,url_for
from utils import QssClient as utl
from utils import TensorClient as tcf
import urllib
import os
app = Flask(__name__)
foo = utl.QssClient()
foo2 = tcf.TensorClient()

@app.route('/')
def api_root():
return 'Welcome'

@app.route('/articles')
def api_articles():
return 'List of ' + url_for('api_articles')

@app.route('/articles/<articleid>')
def api_article(articleid):
return 'You are reading ' + articleid

@app.route('/test1', methods=['GET', 'POST'])
def test1():
resultCode='0'
print (request.method)
if request.method == 'POST':
dic=request.form.to_dict()
print(dic['img'])
foo.baseConvert(dic['img'])
resultCode=foo2.recognize("../pic_data/1.jpg", "../save_bp/lenet5.pb")
#resultCode = '0'
else:
print(request.args.get('img'))
resultCode = '0'
return resultCode
@app.route('/test', methods=['GET', 'POST'])
def test():
resultCode='0'
print (request.method)
if request.method == 'POST':
dic=request.form.to_dict()
print(dic['img'])
foo.baseConvert(dic['img'])
resultCode=foo2.autoCheckImg()
#resultCode = '0'
else:
print(request.args.get('img'))
resultCode = '0'
return resultCode
if __name__ == '__main__':
app.run(host = '0.0.0.0',port = 6001,debug = True)

工具类：
qssclient:

# coding=UTF-8
import sys
import os,base64
import uuid
import requests
class QssClient(object):
def __new__(cls, *args, **kw):
if not hasattr(cls, '_instance'):
orig = super(QssClient, cls)
cls._instance = orig.__new__(cls, *args, **kw)
return cls._instance

def baseConvert(self,filedata):
print ("write ok1")
print filedata
imgdata = base64.b64decode(filedata)
file = open('../pic_data/1.jpg', 'wb')
file.write(imgdata)
print ("write ok2")
file.close()

TensorClient.py:

#coding=utf-8
from PIL import Image, ImageFilter
import tensorflow as tf
import matplotlib as mpl
mpl.use('Agg')
import numpy as np
import matplotlib.pyplot as plt
#import matplotlib.pyplot as plt
import cv2
from skimage import io, transform

class TensorClient(object):
def __new__(cls, *args, **kw):
if not hasattr(cls, '_instance'):
orig = super(TensorClient, cls)
cls._instance = orig.__new__(cls, *args, **kw)
return cls._instance
def imageprepare(self):
file_name = '../pic_data/1.jpg'  # 导入自己的图片地址27  For 5000次训练,20000次以上可以达到99%
# file_name = 'pic_data2/0.png'  # 导入自己的图片地址
# in terminal 'mogrify -format png *.jpg' convert jpg to png
im = Image.open(file_name).convert('L')
im.save("../pic_data/sample.png")
#plt.imshow(im)
#plt.show()
tv = list(im.getdata())  # get pixel values
# normalize pixels to 0 and 1. 0 is pure white, 1 is pure black.
tva = [(255 - x) * 1.0 / 255.0 for x in tv]
# print(tva)
return tva

def weight_variable(self,shape):
initial = tf.truncated_normal(shape, stddev=0.1)
return tf.Variable(initial)

def bias_variable(self,shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)

def conv2d(self,x, W):
return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')

def max_pool_2x2(self,x):
return tf.nn.max_pool(x, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')

#此方法每次执行时要重起服务，不知为什么
def autoCheckImg(self):
result = self.imageprepare()
x = tf.placeholder(tf.float32, [None, 784])
#x = tf.placeholder(tf.float32, [1, 784])
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))

W_conv1 = self.weight_variable([5, 5, 1, 32])
b_conv1 = self.bias_variable([32])

x_image = tf.reshape(x, [-1, 28, 28, 1])
h_conv1 = tf.nn.relu(self.conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = self.max_pool_2x2(h_conv1)

W_conv2 = self.weight_variable([5, 5, 32, 64])
b_conv2 = self.bias_variable([64])

h_conv2 = tf.nn.relu(self.conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 =self. max_pool_2x2(h_conv2)

W_fc1 = self.weight_variable([7 * 7 * 64, 1024])
b_fc1 = self.bias_variable([1024])

h_pool2_flat = tf.reshape(h_pool2, [-1, 7 * 7 * 64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

keep_prob = tf.placeholder(tf.float32)
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

W_fc2 = self.weight_variable([1024, 10])
b_fc2 = self.bias_variable([10])
y_conv = tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)
init_op = tf.global_variables_initializer()
saver = tf.train.Saver()

#saver = tf.train.import_meta_graph("../learning20000/model.ckpt.meta")

checkRlt=0;
with tf.Session() as sess:
#旧方式
sess.run(init_op)
saver.restore(sess, "../learning20000/model.ckpt")  # 这里使用了之前保存的模型参数
#另一种方式
#saver.restore(sess, "../learning20000/model.ckpt")
#sess.run(tf.get_default_graph().get_tensor_by_name("add:0"))

prediction = tf.argmax(y_conv, 1)
predint = prediction.eval(feed_dict={x: [result], keep_prob: 1.0}, session=sess)
print(h_conv2)
print('recognize result:')
print(predint[0])
checkRlt=predint[0]
return str(checkRlt)
#这个方法识别率有问题
def recognize(self,img_path, pb_file_path):
with tf.Graph().as_default():
output_graph_def = tf.GraphDef()

with open(pb_file_path, "rb") as f:
output_graph_def.ParseFromString(f.read())
_ = tf.import_graph_def(output_graph_def, name="")

with tf.Session() as sess:
init = tf.global_variables_initializer()
sess.run(init)

input_x = sess.graph.get_tensor_by_name("input:0")
print(input_x)
keep_prob = sess.graph.get_tensor_by_name("keep_prob:0")
print(keep_prob)
out_softmax = sess.graph.get_tensor_by_name("softmax:0")
print(out_softmax)
out_label = sess.graph.get_tensor_by_name("output:0")
print(out_label)

img = Image.open(img_path).convert('L')
img = img.resize((28, 28))
arr = []
pixelmin = float(img.getpixel((0, 0)))
pixelmax = float(img.getpixel((0, 0)))
for i in range(28):
for j in range(28):

if pixelmin > float(img.getpixel((j, i))):
pixelmin = float(img.getpixel((j, i)))
if pixelmax < float(img.getpixel((j, i))):
pixelmax = float(img.getpixel((j, i)))
# print(pixelmin, pixelmax)
for i in range(28):
for j in range(28):
pixel = (float(img.getpixel((j, i))) - pixelmin) / (pixelmax - pixelmin)
arr.append(pixel)

# print(arr)
img_out_softmax = sess.run(out_softmax, feed_dict={input_x: np.reshape(arr, [-1, 784]), keep_prob: 1.0})

print("img_out_softmax:", img_out_softmax)
prediction_labels = np.argmax(img_out_softmax, axis=1)
print("label:", prediction_labels)
return str(prediction_labels[0])

5.测试客户端



关键代码POST请求

public static string ImageHttpPost(string Url, string postDataStr)
{
try
{
//WriteLog(DateTime.Now + " 影像识别Url：" + Url + " postDataStr：" + postDataStr);
postDataStr = postDataStr.Replace("+", "%2B");
HttpWebRequest request = (HttpWebRequest)WebRequest.Create(Url);
request.Method = "POST";
request.Timeout = 10000;
//request.UserAgent = "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.2; .NET CLR 4.0.30319;)";
request.ContentType = "application/x-www-form-urlencoded";
request.ContentLength = postDataStr.Length;
//增加下面两个属性即可
//request.KeepAlive = false;
//request.ProtocolVersion = HttpVersion.Version10;

StreamWriter writer = new StreamWriter(request.GetRequestStream(), Encoding.ASCII);
writer.Write(postDataStr);
writer.Flush();
writer.Close();
writer.Dispose();
//ServicePointManager.SecurityProtocol = SecurityProtocolType.Tls;
//ServicePointManager.SecurityProtocol = (SecurityProtocolType)3072;
ServicePointManager.SecurityProtocol = SecurityProtocolType.Ssl3 | SecurityProtocolType.Tls;
HttpWebResponse response = (HttpWebResponse)request.GetResponse();
string encoding = response.ContentEncoding;
//if (encoding == null || encoding.Length < 1)
//{
//    encoding = "UTF-8"; //默认编码
//}
Stream myResponseStream = response.GetResponseStream();
StreamReader myStreamReader = new StreamReader(myResponseStream, Encoding.GetEncoding("utf-8"));
string retString = myStreamReader.ReadToEnd();
myStreamReader.Close();
myResponseStream.Close();
return retString;
}
catch (Exception ex)
{
Console.WriteLine(ex);
return null;
}
}

图片生成base64:

/// <summary>
/// 图片生成64
/// </summary>
/// <param name="Imagefilename"></param>
/// <returns></returns>
protected string ImgToBase64String(string Imagefilename)
{
try
{
//生成base64
Bitmap bmp = new Bitmap(Imagefilename);

MemoryStream ms = new MemoryStream();
bmp.Save(ms, System.Drawing.Imaging.ImageFormat.Jpeg);
byte[] arr = new byte[ms.Length];
ms.Position = 0;
ms.Read(arr, 0, (int)ms.Length);
ms.Close();

return Convert.ToBase64String(arr);
}
catch (Exception ex)
{
return null;
}
}

请求API:

//MessageBox.Show("保存成功！");
var base64img = ImgToBase64String(filestring);
// MessageBox.Show("图片准备成功！");
//post
var value = ImageHttpPost("http://192.168.1.168:6001/test", "img=" + base64img);

label3.Text = "识别结束";
if (value == null)
{
label2.Text = "未识别";
}
else
{
label2.Text = value;
}

这个是客户端功能是左则手写0~9后点击保存即可调用服务API进行识别

***************以上内容为本人开发测试后结果转载或引用请标注出处，谢谢***************************