tf-record_decoder_reader.py

import tensorflow as tf
train_record = 'tf-records/coco_train.record-00000-of-00001'

def read_and_decode(filename_queue):
    reader = tf.TFRecordReader()
    _, serialized_example = reader.read(filename_queue)
    features = tf.parse_single_example(
        serialized_example,
        # Defaults are not specified since both keys are required.
        features={
            'image/encoded': tf.FixedLenFeature([], tf.string),
            'image/height': tf.FixedLenFeature([], tf.int64),
            'image/width': tf.FixedLenFeature([], tf.int64),
            'image/filename': tf.FixedLenFeature([], tf.string),
            'image/object/bbox/xmin': tf.VarLenFeature(tf.float32),
            'image/object/bbox/xmax': tf.VarLenFeature(tf.float32),
            'image/object/bbox/ymin': tf.VarLenFeature(tf.float32),
            'image/object/bbox/ymax': tf.VarLenFeature(tf.float32),
            'image/object/class/text': tf.VarLenFeature(tf.string)
        })
    image = tf.decode_raw(features['image/encoded'], tf.uint8)
    xmin = tf.cast(features['image/object/bbox/xmin'], tf.float32)
    ymin = tf.cast(features['image/object/bbox/ymin'], tf.float32)
    xmax = tf.cast(features['image/object/bbox/xmax'], tf.float32)
    ymax = tf.cast(features['image/object/bbox/ymax'], tf.float32)
    label = tf.cast(features['image/object/class/text'], tf.string)
    Iheight = tf.cast(features['image/height'], tf.int32)
    Iwidth = tf.cast(features['image/width'], tf.int32)
    return [image,xmin,ymin,xmax,ymax,label,Iheight,Iwidth]

def get_all_records(FILE):
    with tf.Session() as sess:
        filename_queue = tf.train.string_input_producer([ FILE ])
        image = read_and_decode(filename_queue)
        #print(data[0])
        #image = tf.reshape(image, tf.stack([data[6], data[7], 3]))#height,width
        #image.set_shape([640,480,3])
        init_op = tf.initialize_all_variables()
        sess.run(init_op)
        coord = tf.train.Coordinator()
        threads = tf.train.start_queue_runners(coord=coord)
        for i in range(1):
            example = sess.run([image])
            #img = Image.fromarray(example, 'RGB')
            #img.save( "output/" + str(i) + '-train.png')
            print (len(example[0]))
        coord.request_stop()
        coord.join(threads)

get_all_records(train_record)