字符是识别---11--源数据20*20--训练样本0~Z----基于TensorFlow+CNN实现

 

#coding=utf-8

import os
#图像读取库
from PIL import Image
#矩阵运算库
import numpy as np
import tensorflow as tf



# 训练还是测试
train = True #True False
# 数据文件夹
if train:
    data_dir = "data"
else:
    data_dir = "test"
# 模型文件路径
model_path = "model/image_model"


# 从文件夹读取图片和标签到numpy数组中
# 标签信息在文件名中，例如1_40.jpg表示该图片的标签为1
def read_data(data_dir):
    datas = []
    labels = []
    fpaths = []
    for fname in os.listdir(data_dir):
        fpath = os.path.join(data_dir, fname)
        fpaths.append(fpath)
        image = Image.open(fpath)
        print(fpath)
        data = np.array(image) / 255.0
        label = int(fname.split("_")[0])
        #label = fname.split("_")[0]
        datas.append(data)
        labels.append(label)

    datas = np.array(datas)
    labels = np.array(labels)

    print("shape of datas: {}\tshape of labels: {}".format(datas.shape, labels.shape))
    return fpaths, datas, labels


fpaths, datas, labels = read_data(data_dir)

# 计算有多少类图片
num_classes = len(set(labels))


# 定义Placeholder，存放输入和标签
datas_placeholder = tf.placeholder(tf.float32, [None, 20, 20, 3])
labels_placeholder = tf.placeholder(tf.int32, [None])

# 存放DropOut参数的容器，训练时为0.25，测试时为0
dropout_placeholdr = tf.placeholder(tf.float32)

# 定义卷积层, 20个卷积核, 卷积核大小为5，用Relu激活
conv0 = tf.layers.conv2d(datas_placeholder, 20, 5, activation=tf.nn.relu)
# 定义max-pooling层，pooling窗口为2x2，步长为2x2
pool0 = tf.layers.max_pooling2d(conv0, [2, 2], [2, 2])

# 定义卷积层, 40个卷积核, 卷积核大小为4，用Relu激活
conv1 = tf.layers.conv2d(pool0, 40, 4, activation=tf.nn.relu)
# 定义max-pooling层，pooling窗口为2x2，步长为2x2
pool1 = tf.layers.max_pooling2d(conv1, [2, 2], [2, 2])

# 将3维特征转换为1维向量
flatten = tf.layers.flatten(pool1)

# 全连接层，转换为长度为100的特征向量
fc = tf.layers.dense(flatten, 400, activation=tf.nn.relu)

# 加上DropOut，防止过拟合
dropout_fc = tf.layers.dropout(fc, dropout_placeholdr)

# 未激活的输出层
logits = tf.layers.dense(dropout_fc, num_classes)

predicted_labels = tf.arg_max(logits, 1)


# 利用交叉熵定义损失
losses = tf.nn.softmax_cross_entropy_with_logits(
    labels=tf.one_hot(labels_placeholder, num_classes),
    logits=logits
)
# 平均损失
mean_loss = tf.reduce_mean(losses)

# 定义优化器，指定要优化的损失函数
optimizer = tf.train.AdamOptimizer(learning_rate=1e-2).minimize(losses)


# 用于保存和载入模型
saver = tf.train.Saver()

with tf.Session() as sess:

    if train:
        print("训练模式")
        # 如果是训练，初始化参数
        sess.run(tf.global_variables_initializer())
        # 定义输入和Label以填充容器，训练时dropout为0.25
        train_feed_dict = {
            datas_placeholder: datas,
            labels_placeholder: labels,
            dropout_placeholdr: 0.25
        }
        for step in range(300):#150
            _, mean_loss_val = sess.run([optimizer, mean_loss], feed_dict=train_feed_dict)

            if step % 10 == 0:
                print("step = {}\tmean loss = {}".format(step, mean_loss_val))
        saver.save(sess, model_path)
        print("训练结束，保存模型到{}".format(model_path))
    else:
        print("测试模式")
        # 如果是测试，载入参数
        saver.restore(sess, model_path)
        print("从{}载入模型".format(model_path))
        # label和名称的对照关系
        label_name_dict = {
            0: "0",
            1: "1",
            2: "2",
            3: "3",
            4: "4",
            5: "5",
            6: "6",
            7: "7",
            8: "8",
            9: "9",
            10: "A",
            11: "B",
            12: "C",
            13: "D",
            14: "E",
            15: "F",
            16: "G",
            17: "H",
            18: "I",
            19: "J",
            20: "K",
            21: "L",
            22: "M",
            23: "N",
            24: "P",
            25: "Q",
            26: "R",
            27: "S",
            28: "T",
            29: "U",
            30: "V",
            31: "W",
            32: "X",
            33: "Y",
            34: "Z"
        }
        # 定义输入和Label以填充容器，测试时dropout为0
        test_feed_dict = {
            datas_placeholder: datas,
            labels_placeholder: labels,
            dropout_placeholdr: 0
        }
        predicted_labels_val = sess.run(predicted_labels, feed_dict=test_feed_dict)
        len = predicted_labels_val.shape[0]
        num = 0
        # 真实label与模型预测label
        for fpath, real_label, predicted_label in zip(fpaths, labels, predicted_labels_val):
            # 将label id转换为label名
            real_label_name = label_name_dict[real_label]
            predicted_label_name = label_name_dict[predicted_label]
            print("{}\t{} => {}".format(fpath, real_label_name, predicted_label_name))
            if real_label == predicted_label:
                num = num + 1
                print("正确")
            else:
                print("错误***************************************************************************")
        print("正确率：{}% ，总个数：{}，判断正确个数：{}".format(num / len * 100, len, num))

 

2019-06-26 15:58:18.709062: I tensorflow/core/common_runtime/process_util.cc:69] Creating new thread pool with default inter op setting: 4. Tune using inter_op_parallelism_threads for best performance.
测试模式
从model/image_model载入模型
test\0_1.jpg    0 => D
错误***************************************************************************
test\10_2.jpg    A => 1
错误***************************************************************************
test\11_0.jpg    B => B
正确
test\12_0.jpg    C => C
正确
test\13_0.jpg    D => D
正确
test\14_0.jpg    E => E
正确
test\15_0.jpg    F => E
错误***************************************************************************
test\16_0.jpg    G => G
正确
test\17_0.jpg    H => H
正确
test\18_0.jpg    I => I
正确
test\19_0.jpg    J => 2
错误***************************************************************************
test\1_0.jpg    1 => 1
正确
test\20_1.jpg    K => K
正确
test\21_0.jpg    L => L
正确
test\22_0.jpg    M => H
错误***************************************************************************
test\22_1.jpg    M => F
错误***************************************************************************
test\23_0.jpg    N => W
错误***************************************************************************
test\24_0.jpg    P => P
正确
test\25_0.jpg    Q => Q
正确
test\26_0.jpg    R => F
错误***************************************************************************
test\27_0.jpg    S => S
正确
test\28_0.jpg    T => T
正确
test\29_0.jpg    U => U
正确
test\2_1.jpg    2 => 2
正确
test\30_0.jpg    V => W
错误***************************************************************************
test\31_0.jpg    W => W
正确
test\32_0.jpg    X => X
正确
test\33_0.jpg    Y => Y
正确
test\34_0.jpg    Z => Z
正确
test\3_5.jpg    3 => 3
正确
test\4_0.jpg    4 => 4
正确
test\4_17.jpg    4 => 4
正确
test\5_2.jpg    5 => 5
正确
test\6_1.jpg    6 => 6
正确
test\7_1.jpg    7 => 7
正确
test\8_2.jpg    8 => 8
正确
test\8_4.jpg    8 => 8
正确
test\9_12.jpg    9 => 9
正确
正确率：76.31578947368422% ，总个数：38，判断正确个数：29