tensorflow——TFReord使用图像实例

import os
import glob
from datetime import datetime
import cv2
import  numpy as np
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import tensorflow as tf

image_path='./img/'
images=glob.glob(image_path+'*.jpg')
print(images)



image_labels={
    'dog':0,
    'kangaroo':1
}
#制作tfrecord
def _bytes_feature(value):
    """Returns a bytes_list from a string / byte."""
    if isinstance(value, type(tf.constant(0))):
        # BytesList won't unpack a string from an EagerTensor.
        value = value.numpy()
    return tf.train.Feature(bytes_list=tf.train.BytesList(value=[value]))

def _float_feature(value):
    """Returns a float_list from a float / double."""
    return tf.train.Feature(float_list=tf.train.FloatList(value=[value]))

def _int64_feature(value):
    """Returns an int64_list from a bool / enum / int / uint."""
    return tf.train.Feature(int64_list=tf.train.Int64List(value=[value]))
def image_example(image_string,label):
    """
  Creates a tf.train.Example message ready to be written to a file.
  """
    # Create a dictionary mapping the feature name to the tf.train.Example-compatible
    image_shape=tf.image.decode_jpeg(image_string).shape
    feature = {
        'height': _int64_feature(image_shape[0]),
        'width': _int64_feature(image_shape[1]),
        'depth': _int64_feature(image_shape[2]),
        'label':_int64_feature(label),
        'iamge_raw': _bytes_feature(image_string)
    }
    # Create a Features message using tf.train.Example.
    example_proto = tf.train.Example(features=tf.train.Features(
        feature=feature))
    return example_proto




record_file='image_tfrecord'
counter=0

with tf.io.TFRecordWriter(record_file) as  writer:
    for fname in images:
        with open(fname,'rb') as f:
            image_bgr=f.read()
            image_shape = tf.image.decode_jpeg(image_bgr).shape
            #print(os.path.basename(fname))
            label=int(os.path.basename(fname).replace('.jpg',''))
            tf_example=image_example(image_string=image_bgr,label=label)
            writer.write(tf_example.SerializeToString())

            counter+=1
            print('process {:d} of {:d} images.'.format(counter,len(images)))


#加载创造好的tfrecord
dataset=tf.data.TFRecordDataset('image_tfrecord')
print(dataset)

#example数据都进行了序列化，还需要解析以前写的序列化的string
#解析的格式需要和之前创建的example对应起来
image_feature_description={
    'height': tf.io.FixedLenFeature([],tf.int64),
    'width':tf.io.FixedLenFeature([],tf.int64),
    'depth': tf.io.FixedLenFeature([],tf.int64),
    'label':tf.io.FixedLenFeature([],tf.int64),
    'iamge_raw': tf.io.FixedLenFeature([],tf.string),
}

def parse_tf_example(example_proto):
    #解析出阿里
    parse_example=tf.io.parse_single_example(example_proto,image_feature_description)
    #预处理
    x_train=tf.image.decode_jpeg(parse_example['iamge_raw'],channels=3)
    x_train=tf.image.resize(x_train,(416,416))
    x_train/=255
    label=parse_example['label']
    y_train=label
    return x_train,y_train
train_dataset=dataset.map(parse_tf_example)
print(train_dataset)

#显示数据集
for x in enumerate(train_dataset):
    print(x)