使用TensorFlow的卷积神经网络识别手写数字（3）-识别篇

from PIL import Image
import numpy as np
import tensorflow as tf
import time 


bShowAccuracy = True


# 加载手写图片
def loadHandWritingImage(strFilePath):
    im = Image.open(strFilePath, 'r')
    ndarrayImg = np.array(im.convert("L"), dtype='float')

    return ndarrayImg


# 最大最小值归一化
def normalizeImage(ndarrayImg, maxVal = 255, minVal = 0):
    ndarrayImg = (ndarrayImg - minVal) / (maxVal - minVal)
    return ndarrayImg



# 1)构造自己的手写图片集合，用加载的已训练好的模型识别
print('构造待识别数据...')

# 待识别的手写图片，文件名是0...39
fileList = range(0, 39+1) 

ndarrayImgs = np.zeros((len(fileList), 784)) # x行784列

for index in range(len(fileList)):
    
    # 加载图片
    ndarrayImg = loadHandWritingImage('28-pixel-numbers/' + str(index) + '.png')

    # 归一化
    normalizeImage(ndarrayImg)

    # 转为1x784的数组
    ndarrayImg = ndarrayImg.reshape((1, 784))

    # 加入到测试集中
    ndarrayImgs[index] = ndarrayImg


##import sys
##sys.exit()

# 构建测试样本的实际值集合，用于计算正确率

# 真实结果，用于测试准确度。40行10列
ndarrayLabels = np.eye(10, k=0, dtype='float')
ndarrayLabels = np.vstack((ndarrayLabels, ndarrayLabels))
ndarrayLabels = np.vstack((ndarrayLabels, ndarrayLabels))


# 2)下面开始CNN相关

print('定义Tensor...')

#定义变量和计算公式

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')

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)


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


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

correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y_, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))


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

    # 3)创建saver对象并加载模型
    print('加载已训练好的CNN模型...')
    saver = tf.train.Saver()
    saver.restore(sess, "saved_model/cnn_handwrite_number.ckpt")

    # 测试耗时
    print('进行预测:')
  
    start = time.time()

    # 4）执行预测
    output = sess.run(y_conv, feed_dict={x: ndarrayImgs, keep_prob:1.0})  

    end = time.time()

    print('预测数字为：
', output.argmax(axis=1)) # axis：0表示按列，1表示按行
    print('实际数字为：
', ndarrayLabels.argmax(axis=1))


    if(bShowAccuracy):      
        accu = accuracy.eval(feed_dict={x: ndarrayImgs, y_: ndarrayLabels, keep_prob: 1.0})
        print('识别HateMath苍劲有力的手写数据%d个, 准确率为 %.2f%%, 每个耗时%.5f秒' %
              (len(ndarrayImgs), accu*100, (end-start)/len(ndarrayImgs)))

        

# todo
# 图像分割的准确度