這裡解析所有程式碼 並加入詳細註釋

注意事項：

FCN.py

# coding=utf-8
from __future__ import print_function
import tensorflow as tf
import numpy as np

import TensorflowUtils as utils
import read_MITSceneParsingData as scene_parsing
import datetime
import BatchDatsetReader as dataset
from six.moves import xrange

# 引數設定
FLAGS = tf.flags.FLAGS
tf.flags.DEFINE_integer("batch_size", "2", "batch size for training")
tf.flags.DEFINE_string("logs_dir", "logs/", "path to logs directory")
tf.flags.DEFINE_string("data_dir", "Data_zoo/MIT_SceneParsing/", "path to dataset")
tf.flags.DEFINE_float("learning_rate", "1e-6", "Learning rate for Adam Optimizer")
tf.flags.DEFINE_string("model_dir", "Model_zoo/", "Path to vgg model mat")
tf.flags.DEFINE_bool('debug', "True", "Debug mode: True/ False")
tf.flags.DEFINE_string('mode', "train", "Mode train/ test/ visualize")

MODEL_URL = 'http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-verydeep-19.mat'

MAX_ITERATION = 20000        # 迭代次數
NUM_OF_CLASSESS = 151                # 類別數 151
IMAGE_SIZE = 224                    # 圖片大小 224
fine_tuning = False

# VGG網路部分，weights是權重集合， image是預測影象的向量
def vgg_net(weights, image):
    # VGG網路前五大部分
    layers = (
        'conv1_1', 'relu1_1', 'conv1_2', 'relu1_2', 'pool1',

        'conv2_1', 'relu2_1', 'conv2_2', 'relu2_2', 'pool2',

        'conv3_1', 'relu3_1', 'conv3_2', 'relu3_2', 'conv3_3',
        'relu3_3', 'conv3_4', 'relu3_4', 'pool3',

        'conv4_1', 'relu4_1', 'conv4_2', 'relu4_2', 'conv4_3',
        'relu4_3', 'conv4_4', 'relu4_4', 'pool4',

        'conv5_1', 'relu5_1', 'conv5_2', 'relu5_2', 'conv5_3',
        'relu5_3', 'conv5_4', 'relu5_4'
    )

    net = {}
    current = image     # 預測影象
    for i, name in enumerate(layers):
        kind = name[:4]
        if kind == 'conv':
            kernels, bias = weights[i][0][0][0][0]
            # matconvnet: weights are [width, height, in_channels, out_channels]
            # tensorflow: weights are [height, width, in_channels, out_channels]
            kernels = utils.get_variable(np.transpose(kernels, (1, 0, 2, 3)), name=name + "_w")     # conv1_1_w
            bias = utils.get_variable(bias.reshape(-1), name=name + "_b")       # conv1_1_b
            current = utils.conv2d_basic(current, kernels, bias)        # 前向傳播結果 current
        elif kind == 'relu':
            current = tf.nn.relu(current, name=name)    # relu1_1
            if FLAGS.debug:     # 是否開啟debug模式 true / false
                utils.add_activation_summary(current)       # 畫圖
        elif kind == 'pool':
            # vgg 的前5層的stride都是2，也就是前5層的size依次減小1倍
            # 這裡處理了前4層的stride，用的是平均池化
            # 第5層的pool在下文的外部處理了，用的是最大池化
            # pool1 size縮小2倍
            # pool2 size縮小4倍
            # pool3 size縮小8倍
            # pool4 size縮小16倍
            current = utils.avg_pool_2x2(current)
        net[name] = current     # 每層前向傳播結果放在net中， 是一個字典

    return net


# 預測流程，image是輸入影象，keep_prob dropout比例
def inference(image, keep_prob):
    """
    Semantic segmentation network definition    # 語義分割網路定義
    :param image: input image. Should have values in range 0-255
    :param keep_prob:
    :return:
    """
    # 獲取預訓練網路VGG
    print("setting up vgg initialized conv layers ...")
    # model_dir Model_zoo/
    # MODEL_URL 下載VGG19網址
    model_data = utils.get_model_data(FLAGS.model_dir, MODEL_URL)       # 返回VGG19模型中內容

    mean = model_data['normalization'][0][0][0]                         # 獲得影象均值
    mean_pixel = np.mean(mean, axis=(0, 1))                             # RGB

    weights = np.squeeze(model_data['layers'])                          # 壓縮VGG網路中引數，把維度是1的維度去掉 剩下的就是權重

    processed_image = utils.process_image(image, mean_pixel)            # 影象減均值

    with tf.variable_scope("inference"):                                # 命名作用域 是inference
        image_net = vgg_net(weights, processed_image)                   # 傳入權重引數和預測影象，獲得所有層輸出結果
        conv_final_layer = image_net["conv5_3"]                         # 獲得輸出結果

        pool5 = utils.max_pool_2x2(conv_final_layer)                    # /32 縮小32倍

        W6 = utils.weight_variable([7, 7, 512, 4096], name="W6")        # 初始化第6層的w b
        b6 = utils.bias_variable([4096], name="b6")
        conv6 = utils.conv2d_basic(pool5, W6, b6)
        relu6 = tf.nn.relu(conv6, name="relu6")
        if FLAGS.debug:
            utils.add_activation_summary(relu6)
        relu_dropout6 = tf.nn.dropout(relu6, keep_prob=keep_prob)

        W7 = utils.weight_variable([1, 1, 4096, 4096], name="W7")       # 第7層卷積層
        b7 = utils.bias_variable([4096], name="b7")
        conv7 = utils.conv2d_basic(relu_dropout6, W7, b7)
        relu7 = tf.nn.relu(conv7, name="relu7")
        if FLAGS.debug:
            utils.add_activation_summary(relu7)
        relu_dropout7 = tf.nn.dropout(relu7, keep_prob=keep_prob)

        W8 = utils.weight_variable([1, 1, 4096, NUM_OF_CLASSESS], name="W8")
        b8 = utils.bias_variable([NUM_OF_CLASSESS], name="b8")
        conv8 = utils.conv2d_basic(relu_dropout7, W8, b8)               # 第8層卷積層 分類151類
        # annotation_pred1 = tf.argmax(conv8, dimension=3, name="prediction1")

        # now to upscale to actual image size
        deconv_shape1 = image_net["pool4"].get_shape()                  # 將pool4 1/16結果尺寸拿出來 做融合 [b,h,w,c]
        # 定義反捲積層的 W，B [H, W, OUTC, INC]  輸出個數為pool4層通道個數，輸入為conv8通道個數
        # 擴大兩倍  所以stride = 2  kernel_size = 4
        W_t1 = utils.weight_variable([4, 4, deconv_shape1[3].value, NUM_OF_CLASSESS], name="W_t1")
        b_t1 = utils.bias_variable([deconv_shape1[3].value], name="b_t1")
        # 輸入為conv8特徵圖，使得其特徵圖大小擴大兩倍，並且特徵圖個數變為pool4的通道數
        conv_t1 = utils.conv2d_transpose_strided(conv8, W_t1, b_t1, output_shape=tf.shape(image_net["pool4"]))
        fuse_1 = tf.add(conv_t1, image_net["pool4"], name="fuse_1")     # 進行融合 逐畫素相加

        # 獲得pool3尺寸 是原圖大小的1/8
        deconv_shape2 = image_net["pool3"].get_shape()
        # 輸出通道數為pool3通道數，  輸入通道數為pool4通道數
        W_t2 = utils.weight_variable([4, 4, deconv_shape2[3].value, deconv_shape1[3].value], name="W_t2")
        b_t2 = utils.bias_variable([deconv_shape2[3].value], name="b_t2")
        # 將上一層融合結果fuse_1在擴大兩倍，輸出尺寸和pool3相同
        conv_t2 = utils.conv2d_transpose_strided(fuse_1, W_t2, b_t2, output_shape=tf.shape(image_net["pool3"]))
        # 融合操作deconv(fuse_1) + pool3
        fuse_2 = tf.add(conv_t2, image_net["pool3"], name="fuse_2")

        shape = tf.shape(image)     # 獲得原始影象大小
        # 堆疊列表，反捲積輸出尺寸，[b，原圖H，原圖W，類別個數]
        deconv_shape3 = tf.stack([shape[0], shape[1], shape[2], NUM_OF_CLASSESS])
        # 建立反捲積w[8倍擴大需要ks=16, 輸出通道數為類別個數， 輸入通道數pool3通道數]
        W_t3 = utils.weight_variable([16, 16, NUM_OF_CLASSESS, deconv_shape2[3].value], name="W_t3")
        b_t3 = utils.bias_variable([NUM_OF_CLASSESS], name="b_t3")
        # 反捲積，fuse_2反捲積，輸出尺寸為 [b，原圖H，原圖W，類別個數]
        conv_t3 = utils.conv2d_transpose_strided(fuse_2, W_t3, b_t3, output_shape=deconv_shape3, stride=8)

        # 目前conv_t3的形式為size為和原始影象相同的size，通道數與分類數相同
        # 這句我的理解是對於每個畫素位置，根據第3維度（通道數）通過argmax能計算出這個畫素點屬於哪個分類
        # 也就是對於每個畫素而言，NUM_OF_CLASSESS個通道中哪個數值最大，這個畫素就屬於哪個分類
        # 每個畫素點有21個值，哪個值最大就屬於那一類
        # 返回一張圖，每一個點對於其來別資訊shape=[b,h,w]
        annotation_pred = tf.argmax(conv_t3, dimension=3, name="prediction")
    # 從第三維度擴充套件 形成[b,h,w,c] 其中c=1, conv_t3最後具有21深度的特徵圖
    return tf.expand_dims(annotation_pred, dim=3), conv_t3


def train(loss_val, var_list):
    """

    :param loss_val:  損失函式
    :param var_list:  需要優化的值
    :return:
    """
    optimizer = tf.train.AdamOptimizer(FLAGS.learning_rate)
    grads = optimizer.compute_gradients(loss_val, var_list=var_list)
    if FLAGS.debug:
        # print(len(var_list))
        for grad, var in grads:
            utils.add_gradient_summary(grad, var)
    return optimizer.apply_gradients(grads)     # 返回迭代梯度


def main(argv=None):
    # dropout保留率
    keep_probability = tf.placeholder(tf.float32, name="keep_probabilty")
    # 影象佔坑
    image = tf.placeholder(tf.float32, shape=[None, IMAGE_SIZE, IMAGE_SIZE, 3], name="input_image")
    # 標籤佔坑
    annotation = tf.placeholder(tf.int32, shape=[None, IMAGE_SIZE, IMAGE_SIZE, 1], name="annotation")

    # 預測一個batch影象  獲得預測圖[b,h,w,c=1]  結果特徵圖[b,h,w,c=151]
    pred_annotation, logits = inference(image, keep_probability)
    tf.summary.image("input_image", image, max_outputs=2)
    tf.summary.image("ground_truth", tf.cast(annotation, tf.uint8), max_outputs=2)
    tf.summary.image("pred_annotation", tf.cast(pred_annotation, tf.uint8), max_outputs=2)
    # 空間交叉熵損失函式[b,h,w,c=151]  和labels[b,h,w]    每一張圖分別對比
    loss = tf.reduce_mean((tf.nn.sparse_softmax_cross_entropy_with_logits(logits=logits,
                                                                          labels=tf.squeeze(annotation, squeeze_dims=[3]),
                                                                          name="entropy")))
    tf.summary.scalar("entropy", loss)

    # 返回需要訓練的變數列表
    trainable_var = tf.trainable_variables()
    if FLAGS.debug:
        for var in trainable_var:
            utils.add_to_regularization_and_summary(var)

    # 傳入損失函式和需要訓練的變數列表
    train_op = train(loss, trainable_var)

    print("Setting up summary op...")
    # 生成繪圖資料
    summary_op = tf.summary.merge_all()

    print("Setting up image reader...")
    # data_dir = Data_zoo/MIT_SceneParsing/
    # training: [{image: 圖片全路徑， annotation:標籤全路徑， filename:圖片名字}] [{}][{}]
    train_records, valid_records = scene_parsing.read_dataset(FLAGS.data_dir)
    print(len(train_records))   # 長度
    print(len(valid_records))

    print("Setting up dataset reader")
    image_options = {'resize': True, 'resize_size': IMAGE_SIZE}
    if FLAGS.mode == 'train':
        # 讀取圖片 產生類物件 其中包含所有圖片資訊
        train_dataset_reader = dataset.BatchDatset(train_records, image_options)
    validation_dataset_reader = dataset.BatchDatset(valid_records, image_options)

    sess = tf.Session()

    print("Setting up Saver...")
    saver = tf.train.Saver()
    summary_writer = tf.summary.FileWriter(FLAGS.logs_dir, sess.graph)

    sess.run(tf.global_variables_initializer())
    # logs/
    if fine_tuning:
        ckpt = tf.train.get_checkpoint_state(FLAGS.logs_dir)    # 訓練斷點回復
        if ckpt and ckpt.model_checkpoint_path:                 # 如果存在checkpoint檔案 則恢復sess
            saver.restore(sess, ckpt.model_checkpoint_path)
            print("Model restored...")

    if FLAGS.mode == "train":
        for itr in range(MAX_ITERATION):
            # 讀取下一batch
            train_images, train_annotations = train_dataset_reader.next_batch(FLAGS.batch_size)
            feed_dict = {image: train_images, annotation: train_annotations, keep_probability: 0.85}

            # 迭代優化需要訓練的變數
            sess.run(train_op, feed_dict=feed_dict)

            if itr % 10 == 0:
                # 迭代10次列印顯示
                train_loss, summary_str = sess.run([loss, summary_op], feed_dict=feed_dict)
                print("Step: %d, Train_loss:%g" % (itr, train_loss))
                summary_writer.add_summary(summary_str, itr)

            if itr % 500 == 0:
                # 迭代500 次驗證
                valid_images, valid_annotations = validation_dataset_reader.next_batch(FLAGS.batch_size)
                valid_loss = sess.run(loss, feed_dict={image: valid_images, annotation: valid_annotations,
                                                       keep_probability: 1.0})
                print("%s ---> Validation_loss: %g" % (datetime.datetime.now(), valid_loss))
                # 儲存模型
                saver.save(sess, FLAGS.logs_dir + "model.ckpt", itr)

    elif FLAGS.mode == "visualize":
        # 視覺化
        valid_images, valid_annotations = validation_dataset_reader.get_random_batch(FLAGS.batch_size)
        # pred_annotation預測結果圖
        pred = sess.run(pred_annotation, feed_dict={image: valid_images, annotation: valid_annotations,
                                                    keep_probability: 1.0})
        valid_annotations = np.squeeze(valid_annotations, axis=3)
        pred = np.squeeze(pred, axis=3)

        for itr in range(FLAGS.batch_size):
            utils.save_image(valid_images[itr].astype(np.uint8), FLAGS.logs_dir, name="inp_" + str(5+itr))
            utils.save_image(valid_annotations[itr].astype(np.uint8), FLAGS.logs_dir, name="gt_" + str(5+itr))
            utils.save_image(pred[itr].astype(np.uint8), FLAGS.logs_dir, name="pred_" + str(5+itr))
            print("Saved image: %d" % itr)


if __name__ == "__main__":
    tf.app.run()
 

read_MITSceneParsingData.py

# coding=utf-8
__author__ = 'charlie'
import numpy as np
import os
import random
from six.moves import cPickle as pickle
from tensorflow.python.platform import gfile
import glob

import TensorflowUtils as utils

# DATA_URL = 'http://sceneparsing.csail.mit.edu/data/ADEChallengeData2016.zip'
DATA_URL = 'http://data.csail.mit.edu/places/ADEchallenge/ADEChallengeData2016.zip'


def read_dataset(data_dir):
    # data_dir = Data_zoo / MIT_SceneParsing /
    pickle_filename = "MITSceneParsing.pickle"
    # 檔案路徑  Data_zoo / MIT_SceneParsing / MITSceneParsing.pickle
    pickle_filepath = os.path.join(data_dir, pickle_filename)
    if not os.path.exists(pickle_filepath):
        utils.maybe_download_and_extract(data_dir, DATA_URL, is_zipfile=True)       # 不存在檔案 則下載
        SceneParsing_folder = os.path.splitext(DATA_URL.split("/")[-1])[0]          # ADEChallengeData2016
        # result =   {training: [{image: 圖片全路徑， annotation:標籤全路徑， filename:圖片名字}] [][]
        #            validation:[{image:圖片全路徑， annotation:標籤全路徑， filename:圖片名字}] [] []}
        result = create_image_lists(os.path.join(data_dir, SceneParsing_folder))    # Data_zoo / MIT_SceneParsing / ADEChallengeData2016
        print ("Pickling ...")      # 製作pickle檔案
        with open(pickle_filepath, 'wb') as f:
            pickle.dump(result, f, pickle.HIGHEST_PROTOCOL)
    else:
        print ("Found pickle file!")

    with open(pickle_filepath, 'rb') as f:      # 開啟pickle檔案
        result = pickle.load(f)                 # 讀取
        training_records = result['training']
        validation_records = result['validation']
        del result
    # training: [{image: 圖片全路徑， annotation:標籤全路徑， filename:圖片名字}] [{}][{}]
    return training_records, validation_records


def create_image_lists(image_dir):
    """

    :param image_dir:   Data_zoo / MIT_SceneParsing / ADEChallengeData2016
    :return:
    """
    if not gfile.Exists(image_dir):
        print("Image directory '" + image_dir + "' not found.")
        return None
    directories = ['training', 'validation']
    image_list = {}     # 影象字典   training:[]  validation:[]

    for directory in directories:       # 訓練集和驗證集 分別製作
        file_list = []
        image_list[directory] = []
        # Data_zoo/MIT_SceneParsing/ADEChallengeData2016/images/training/*.jpg
        file_glob = os.path.join(image_dir, "images", directory, '*.' + 'jpg')
        # 加入檔案列表  包含所有圖片檔案全路徑+檔名字  如 Data_zoo/MIT_SceneParsing/ADEChallengeData2016/images/training/hi.jpg
        file_list.extend(glob.glob(file_glob))

        if not file_list:   # 檔案為空
            print('No files found')
        else:
            for f in file_list:     # 掃描檔案列表   這裡f對應檔案全路徑
                # 獲取圖片名字 hi
                filename = os.path.splitext(f.split("/")[-1])[0]
                # Data_zoo/MIT_SceneParsing/ADEChallengeData2016/annotations/training/*.png
                annotation_file = os.path.join(image_dir, "annotations", directory, filename + '.png')
                if os.path.exists(annotation_file):     # 如果檔案路徑存在
                    #  image:圖片全路徑， annotation:標籤全路徑， filename:圖片名字
                    record = {'image': f, 'annotation': annotation_file, 'filename': filename}
                    # image_list{training:[{image:圖片全路徑， annotation:標籤全路徑， filename:圖片名字}] [] []
                    #            validation:[{image:圖片全路徑， annotation:標籤全路徑， filename:圖片名字}] [] []}
                    image_list[directory].append(record)
                else:
                    print("Annotation file not found for %s - Skipping" % filename)
        # 對圖片列表進行洗牌
        random.shuffle(image_list[directory])
        no_of_images = len(image_list[directory])   # 包含圖片檔案的個數
        print ('No. of %s files: %d' % (directory, no_of_images))

    return image_list
 

TensorflowUitls.py

# coding=utf-8
__author__ = 'Charlie'
# Utils used with tensorflow implemetation
import tensorflow as tf
import numpy as np
import scipy.misc as misc
import os, sys
from six.moves import urllib
import tarfile
import zipfile
import scipy.io


# 獲取VGG預訓練模型
def get_model_data(dir_path, model_url):
    # model_dir Model_zoo/
    # MODEL_URL 下載VGG19網址
    maybe_download_and_extract(dir_path, model_url)     # 判斷檔案目錄和檔案是否存在， 不存在則下載
    filename = model_url.split("/")[-1]                 # 將url按/切分， 取最後一個字串作為檔名
    filepath = os.path.join(dir_path, filename)         # dir_path/filename     檔案全路徑
    if not os.path.exists(filepath):                    # 判斷是否存在此檔案
        raise IOError("VGG Model not found!")
    data = scipy.io.loadmat(filepath)                   # 使用io讀取VGG.mat檔案
    return data


def maybe_download_and_extract(dir_path, url_name, is_tarfile=False, is_zipfile=False):
    # dir_path Model_zoo/
    # url_name 下載VGG19網址
    if not os.path.exists(dir_path):        # 判斷檔案路徑是否存在，如果不存在則建立此路徑
        os.makedirs(dir_path)
    filename = url_name.split('/')[-1]      # 將url中 按照/切分，並取最後一個字串 作為檔名字
    filepath = os.path.join(dir_path, filename)     # 檔案路徑 = dir_path/filename
    if not os.path.exists(filepath):         # 判斷此路徑是否存在（此檔案），如果不存在，則下載
        def _progress(count, block_size, total_size):       # 內部函式
            sys.stdout.write(
                '\r>> Downloading %s %.1f%%' % (filename, float(count * block_size) / float(total_size) * 100.0))
            sys.stdout.flush()

        filepath, _ = urllib.request.urlretrieve(url_name, filepath, reporthook=_progress)    # 將url中檔案 下載到filepath路徑中
        print()
        statinfo = os.stat(filepath)
        print('Succesfully downloaded', filename, statinfo.st_size, 'bytes.')
        if is_tarfile:          # 如果是tar檔案， 解壓縮
            tarfile.open(filepath, 'r:gz').extractall(dir_path)
        elif is_zipfile:        # 如果是zip檔案 解壓縮
            with zipfile.ZipFile(filepath) as zf:
                zip_dir = zf.namelist()[0]
                zf.extractall(dir_path)

 

BatchDatsetReader.py

# coding=utf-8
"""
Code ideas from https://github.com/Newmu/dcgan and tensorflow mnist dataset reader
"""
import numpy as np
import scipy.misc as misc


class BatchDatset:
    files = []
    images = []
    annotations = []
    image_options = {}
    batch_offset = 0
    epochs_completed = 0

    def __init__(self, records_list, image_options={}):
        """
        Intialize a generic file reader with batching for list of files
        :param records_list: list of file records to read -
        sample record: {'image': f, 'annotation': annotation_file, 'filename': filename}
        :param image_options: A dictionary of options for modifying the output image
        Available options:
        resize = True/ False
        resize_size = #size of output image - does bilinear resize
        color=True/False
        """
        print("Initializing Batch Dataset Reader...")
        print(image_options)
        self.files = records_list       # 檔案列表
        self.image_options = image_options  # 圖片操作方式 resize  224
        self._read_images()

    def _read_images(self):
        self.__channels = True
        # 掃描files字典中所有image 圖片全路徑
        # 根據檔案全路徑讀取影象，並將其擴充為RGB格式
        self.images = np.array([self._transform(filename['image']) for filename in self.files])
        self.__channels = False

        # 掃描files字典中所有annotation 圖片全路徑
        # 根據檔案全路徑讀取影象，並將其擴充為三通道格式
        self.annotations = np.array(
            [np.expand_dims(self._transform(filename['annotation']), axis=3) for filename in self.files])
        print (self.images.shape)
        print (self.annotations.shape)

    def _transform(self, filename):
        # 讀取檔案圖片
        image = misc.imread(filename)
        if self.__channels and len(image.shape) < 3:  # make sure images are of shape(h,w,3)
            # 將圖片三個通道設定為一樣的圖片
            image = np.array([image for i in range(3)])

        if self.image_options.get("resize", False) and self.image_options["resize"]:

            resize_size = int(self.image_options["resize_size"])
            # 使用最近鄰插值法resize圖片
            resize_image = misc.imresize(image,
                                         [resize_size, resize_size], interp='nearest')
        else:
            resize_image = image

        return np.array(resize_image)       # 返回已經resize的圖片

    def get_records(self):
        """
        返回圖片和標籤全路徑
        :return:
        """
        return self.images, self.annotations

    def reset_batch_offset(self, offset=0):
        """
        剩下的batch
        :param offset:
        :return:
        """
        self.batch_offset = offset

    def next_batch(self, batch_size):
        # 當前第幾個batch
        start = self.batch_offset
        # 讀取下一個batch  所有offset偏移量+batch_size
        self.batch_offset += batch_size
        # iamges儲存所有圖片資訊 images.shape(len, h, w)
        if self.batch_offset > self.images.shape[0]:      # 如果下一個batch的偏移量超過了圖片總數 說明完成了一個epoch
            # Finished epoch
            self.epochs_completed += 1      # epochs完成總數+1
            print("****************** Epochs completed: " + str(self.epochs_completed) + "******************")
            # Shuffle the data
            perm = np.arange(self.images.shape[0])      # arange生成陣列(0 - len-1) 獲取圖片索引
            np.random.shuffle(perm)         # 對圖片索引洗牌
            self.images = self.images[perm]     # 洗牌之後的圖片順序
            self.annotations = self.annotations[perm]
            # Start next epoch
            start = 0           # 下一個epoch從0開始
            self.batch_offset = batch_size  # 已完成的batch偏移量

        end = self.batch_offset             # 開始到結束self.batch_offset   self.batch_offset+batch_size
        return self.images[start:end], self.annotations[start:end]      # 取出batch

    def get_random_batch(self, batch_size):
        # 按照一個batch_size一個塊  進行對所有圖片總數進行隨機操作， 相當於洗牌工作
        indexes = np.random.randint(0, self.images.shape[0], size=[batch_size]).tolist()
        return self.images[indexes], self.annotations[indexes]