本文共 6045 字，大约阅读时间需要 20 分钟。

#coding=utf-8import os#图像读取库from PIL import Image#矩阵运算库import numpy as npimport tensorflow as tf# 训练还是测试train = False #True FalseStep_total = 150# 数据文件夹if train:    data_dir = "train_35" #data train_35else:    data_dir = "test_35"# 模型文件路径model_path = "model35/image_model"  #model  model35# 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1def 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, labelsfpaths, 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0dropout_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2x2pool0 = 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2x2pool1 = 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(Step_total):#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))

测试集如图：

2019-06-26 18:32:27.031947: 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model35/image_model载入模型test_35\0_1.jpg	0 => 0正确test_35\10_2.jpg	A => A正确test_35\11_0.jpg	B => E错误***************************************************************************test_35\12_0.jpg	C => C正确test_35\13_0.jpg	D => D正确test_35\14_0.jpg	E => E正确test_35\15_0.jpg	F => F正确test_35\16_0.jpg	G => G正确test_35\17_0.jpg	H => H正确test_35\18_0.jpg	I => I正确test_35\19_0.jpg	J => J正确test_35\1_0.jpg	1 => 1正确test_35\20_1.jpg	K => K正确test_35\21_0.jpg	L => L正确test_35\22_0.jpg	M => M正确test_35\22_1.jpg	M => M正确test_35\23_0.jpg	N => N正确test_35\24_0.jpg	P => P正确test_35\25_0.jpg	Q => Q正确test_35\26_0.jpg	R => R正确test_35\27_0.jpg	S => S正确test_35\28_0.jpg	T => T正确test_35\29_0.jpg	U => U正确test_35\2_1.jpg	2 => 2正确test_35\30_0.jpg	V => V正确test_35\31_0.jpg	W => W正确test_35\32_0.jpg	X => X正确test_35\33_0.jpg	Y => Y正确test_35\34_0.jpg	Z => Z正确test_35\3_5.jpg	3 => 3正确test_35\4_0.jpg	4 => 4正确test_35\4_1.jpg	4 => 4正确test_35\4_17.jpg	4 => 4正确test_35\5_2.jpg	5 => 5正确test_35\6_1.jpg	6 => 6正确test_35\7_1.jpg	7 => 7正确test_35\8_2.jpg	8 => 8正确test_35\8_4.jpg	8 => 8正确test_35\9_12.jpg	9 => 9正确正确率：97.43589743589743% ，总个数：39，判断正确个数：38Process finished with exit code 0

转载地址：http://elsws.baihongyu.com/