def process_image(self, encoded_image, thread_id=0):
return image_processing.process_image(
encoded_image,
is_training=self.is_training(),
height=self.config.image_height,
width=self.config.image_width,
thread_id=thread_id,
image_format=self.config.image_format)
def decode_images(encoded_images):
decoded_images = []
for image in encoded_images:
decoded_image = image_processing.process_image(image,
is_training=True,
thread_id=0,
height=299,
width=299,
resize_height=346,
resize_width=346,
image_format='jpeg')
decoded_images.append(decoded_image)
return decoded_images
def process_image(self, encoded_image, thread_id=0):
"""Decodes and processes an image string.
Args:
encoded_image: A scalar string Tensor; the encoded image.
thread_id: Preprocessing thread id used to select the ordering of color
distortions.
Returns:
A float32 Tensor of shape [height, width, 3]; the processed image.
"""
return image_processing.process_image(encoded_image,
is_training=self.is_training(),
height=self.config.image_height,
width=self.config.image_width,
thread_id=thread_id,
image_format=self.config.image_format)
def process_image(self, encoded_image, thread_id=0):
"""Decodes and processes an image string.
Args:
encoded_image: A scalar string Tensor; the encoded image.
thread_id: Preprocessing thread id used to select the ordering of color
distortions.
Returns:
A float32 Tensor of shape [height, width, 3]; the processed image.
"""
return image_processing.process_image(encoded_image,
is_training=self.is_training(),
height=self.config.image_height,
width=self.config.image_width,
thread_id=thread_id,
image_format=self.config.image_format)
def build_input(self):
#TFRecord로부터 이미지와 캡션 불러오기
input_queue = self.prefetch_tfrecords()
images_captions_pairs = []
for thread_id in range(tf.flags.FLAGS.num_preprocess_threads):
context, sequence = tf.parse_single_sequence_example(
input_queue.dequeue(),
context_features={
'image/data': tf.FixedLenFeature([], dtype=tf.string)
},
sequence_features={
'image/caption_ids':
tf.FixedLenSequenceFeature([], dtype=tf.int64),
})
encoded_image = context['image/data']
#captions = sequence['image/caption_ids']
captions = sequence.values()
# 이미지 디코딩 하기
decoded_image = image_processing.process_image(
encoded_image,
is_training=True,
thread_id=thread_id,
height=tf.flags.FLAGS.image_height,
width=tf.flags.FLAGS.image_width,
resize_height=tf.flags.FLAGS.resize_height,
resize_width=tf.flags.FLAGS.resize_width,
image_format=tf.flags.FLAGS.image_format)
images_captions_pairs.append([decoded_image, captions])
skip_gram_pairs = []
for image, captions in images_captions_pairs:
for word in captions:
skip_gram_pairs.append((image, word))
#skip_gram_pairs = make_skip_gram_for_image.make_skip_gram_for_image(images_captions_pairs)
# 배치 입력 만들기
self.batch_images, self.batch_contexts = make_batch_input.make_batch_input(
skip_gram_pairs)
self.image_embeddings = make_image_embeddings_cnn.make_image_embeddings_cnn(
self.batch_images, self.voc_size)
