Pass:

Handwritten digit recognition -- a detailed explanation of the official case of convolutional neural network model (based on Tensorflow,Python)

Handwritten digit recognition -- a detailed explanation of the official case of Softmax regression model (based on Tensorflow,Python)

After running the programFour documents, again by handImage judgment recognition probability

Code:

import numpy as np
import tensorflow as tf
from flask import Flask, jsonify, render_template, request
import numpy as np
from PIL import Image
from mnist import model
import matplotlib.pyplot as plt
from matplotlib.image import imread

# tf.placeholder(dtype, shape=None, name=None)
# This function can be understood as a parameter, which is used to define the process and assign specific
# dtype: Data type. What is commonly used is tf.float32,tf.float64 Equal value type
# shape: Data shape. The default is None，It can be a one-dimensional value or a multi-dimensional value, such as[2,3], [None, 3]Indicates that the column is 3 and the row is indefinite
# name: Name.
# Return: Tensor type

x = tf.placeholder("float", [None, 784])

'''Used for running TensorFlow The class of the operation. '''
# session Possible resources, such as: tf.Variable，tf.QueueBase and tf.ReaderBase. 
# It is very important to release these resources when they are no longer needed.
# To do this, please session Call in tf.Session.close Method, or use session As context manager
sess = tf.Session()

# Both save and restore need to instantiate a tf.train.Saver. 
# saver = tf.train.Saver()
# In the training cycle, call regularly saver.save() Method, write to the folder the checkpoint Papers.
# saver.save(sess, FLAGS.train_dir, global_step=step)
# After that, you can use saver.restore() Method, overloads the parameters of the model, continues training or is used to test the data.
# saver.restore(sess, FLAGS.train_dir)

# restore trained data
with tf.variable_scope("regression"):
    y1, variables = model.regression(x)
saver = tf.train.Saver(variables)
saver.restore(sess, "mnist/data/regression.ckpt")


# tf.get_variable(<name>, <shape>, <initializer>) Creates or returns a variable of the given name
# tf.variable_scope(<scope_name>) Management to get_variable()The scope of the variable name of
with tf.variable_scope("convolutional"):
    keep_prob = tf.placeholder("float")
    y2, variables = model.convolutional(x, keep_prob)
saver = tf.train.Saver(variables)
saver.restore(sess, "mnist/data/convolutional.ckpt")


def regression(input):
    # print('-------------------regression')
    # print('y2:' + str(y1))
    # print(input)
    return sess.run(y1, feed_dict={x: input}).flatten().tolist()

# run(
#     fetches,
#     feed_dict=None,
#     options=None,
#     run_metadata=None
# )
def convolutional(input):
    # print('-------------------convolutional')
    # print('y2:' + str(y2))
    # print( input)
    return sess.run(y2, feed_dict={x: input, keep_prob: 1.0}).flatten().tolist()


# im = Image.open(r'C:\Users\admin\Desktop\No title.png')
# im2 = np.array(im)
# print(im2)

# img = imread(r'C:\Users\admin\Desktop\No title.png')  # Read in the image (set the appropriate path!)
# plt.imshow(img)
# plt.arr
# plt.show()

# Read pictures
im = Image.open(r'C:\Users\admin\Desktop\2.png')
# display picture
# im.show()
im = im.convert("L")
# im.show()
data = im.getdata()
data = np.matrix(data)

 #     print data
# Transform to 512*512
data = np.reshape(data, (784, 1))
# new_im = Image.fromarray(data)
# # display picture
# new_im.show()
input = ((255 - np.array(data, dtype=np.uint8)) / 255.0).reshape(1, 784)

# # print(input)
output1 = regression(input)
output2 = convolutional(input)
print(output1)
print(output2)

Handwritten digit recognition