tensorflow官方文档学习 Image Recognition（Inception v3模型）

# Copyright 2015 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================


"""Simple image classification with Inception.


Run image classification with Inception trained on ImageNet 2012 Challenge data
set.


This program creates a graph from a saved GraphDef protocol buffer,
and runs inference on an input JPEG image. It outputs human readable
strings of the top 5 predictions along with their probabilities.


Change the --image_file argument to any jpg image to compute a
classification of that image.


Please see the tutorial and website for a detailed description of how
to use this script to perform image recognition.


https://tensorflow.org/tutorials/image_recognition/
"""


from __future__ import absolute_import
from __future__ import division
from __future__ import print_function


import argparse
import os.path
import re
import sys
import tarfile


import numpy as np
from six.moves import urllib
import tensorflow as tf


FLAGS = None


# pylint: disable=line-too-long
DATA_URL = 'http://download.tensorflow.org/models/image/imagenet/inception-2015-12-05.tgz'
# pylint: enable=line-too-long




class NodeLookup(object):
  """Converts integer node ID's to human readable labels."""


  def __init__(self,
               label_lookup_path=None,
               uid_lookup_path=None):
    if not label_lookup_path:
      label_lookup_path = os.path.join(
          FLAGS.model_dir, 'imagenet_2012_challenge_label_map_proto.pbtxt')
    if not uid_lookup_path:
      uid_lookup_path = os.path.join(
          FLAGS.model_dir, 'imagenet_synset_to_human_label_map.txt')
    self.node_lookup = self.load(label_lookup_path, uid_lookup_path)


  def load(self, label_lookup_path, uid_lookup_path):
    """Loads a human readable English name for each softmax node.


    Args:
      label_lookup_path: string UID to integer node ID.
      uid_lookup_path: string UID to human-readable string.


    Returns:
      dict from integer node ID to human-readable string.
    """
    if not tf.gfile.Exists(uid_lookup_path):
      tf.logging.fatal('File does not exist %s', uid_lookup_path)
    if not tf.gfile.Exists(label_lookup_path):
      tf.logging.fatal('File does not exist %s', label_lookup_path)


    # Loads mapping from string UID to human-readable string
    proto_as_ascii_lines = tf.gfile.GFile(uid_lookup_path).readlines()
    uid_to_human = {}
    p = re.compile(r'[n\d]*[ \S,]*')
    for line in proto_as_ascii_lines:
      parsed_items = p.findall(line)
      uid = parsed_items[0]
      human_string = parsed_items[2]
      uid_to_human[uid] = human_string


    # Loads mapping from string UID to integer node ID.
    node_id_to_uid = {}
    proto_as_ascii = tf.gfile.GFile(label_lookup_path).readlines()
    for line in proto_as_ascii:
      if line.startswith('  target_class:'):
        target_class = int(line.split(': ')[1])
      if line.startswith('  target_class_string:'):
        target_class_string = line.split(': ')[1]
        node_id_to_uid[target_class] = target_class_string[1:-2]


    # Loads the final mapping of integer node ID to human-readable string
    node_id_to_name = {}
    for key, val in node_id_to_uid.items():
      if val not in uid_to_human:
        tf.logging.fatal('Failed to locate: %s', val)
      name = uid_to_human[val]
      node_id_to_name[key] = name


    return node_id_to_name


  def id_to_string(self, node_id):
    if node_id not in self.node_lookup:
      return ''
    return self.node_lookup[node_id]




def create_graph():
  """Creates a graph from saved GraphDef file and returns a saver."""
  # Creates graph from saved graph_def.pb.
  with tf.gfile.FastGFile(os.path.join(
      FLAGS.model_dir, 'classify_image_graph_def.pb'), 'rb') as f:
    graph_def = tf.GraphDef()
    graph_def.ParseFromString(f.read())
    _ = tf.import_graph_def(graph_def, name='')




def run_inference_on_image(image):
  """Runs inference on an image.


  Args:
    image: Image file name.


  Returns:
    Nothing
  """
  if not tf.gfile.Exists(image):
    tf.logging.fatal('File does not exist %s', image)
  image_data = tf.gfile.FastGFile(image, 'rb').read()


  # Creates graph from saved GraphDef.
  create_graph()


  with tf.Session() as sess:
    # Some useful tensors:
    # 'softmax:0': A tensor containing the normalized prediction across
    #   1000 labels.
    # 'pool_3:0': A tensor containing the next-to-last layer containing 2048
    #   float description of the image.
    # 'DecodeJpeg/contents:0': A tensor containing a string providing JPEG
    #   encoding of the image.
    # Runs the softmax tensor by feeding the image_data as input to the graph.
    softmax_tensor = sess.graph.get_tensor_by_name('softmax:0')
    predictions = sess.run(softmax_tensor,
                           {'DecodeJpeg/contents:0': image_data})
    predictions = np.squeeze(predictions)


    # Creates node ID --> English string lookup.
    node_lookup = NodeLookup()


    top_k = predictions.argsort()[-FLAGS.num_top_predictions:][::-1]
    for node_id in top_k:
      human_string = node_lookup.id_to_string(node_id)
      score = predictions[node_id]
      print('%s (score = %.5f)' % (human_string, score))




def maybe_download_and_extract():
  """Download and extract model tar file."""
  dest_directory = FLAGS.model_dir
  if not os.path.exists(dest_directory):
    os.makedirs(dest_directory)
  filename = DATA_URL.split('/')[-1]
  filepath = os.path.join(dest_directory, filename)
  if not os.path.exists(filepath):
    def _progress(count, block_size, total_size):
      sys.stdout.write('\r>> Downloading %s %.1f%%' % (
          filename, float(count * block_size) / float(total_size) * 100.0))
      sys.stdout.flush()
    filepath, _ = urllib.request.urlretrieve(DATA_URL, filepath, _progress)
    print()
    statinfo = os.stat(filepath)
    print('Successfully downloaded', filename, statinfo.st_size, 'bytes.')
  tarfile.open(filepath, 'r:gz').extractall(dest_directory)




def main(_):
  maybe_download_and_extract()
  image = (FLAGS.image_file if FLAGS.image_file else
           os.path.join(FLAGS.model_dir, 'cropped_panda.jpg'))
  run_inference_on_image(image)




if __name__ == '__main__':
  parser = argparse.ArgumentParser()
  # classify_image_graph_def.pb:
  #   Binary representation of the GraphDef protocol buffer.
  # imagenet_synset_to_human_label_map.txt:
  #   Map from synset ID to a human readable string.
  # imagenet_2012_challenge_label_map_proto.pbtxt:
  #   Text representation of a protocol buffer mapping a label to synset ID.
  parser.add_argument(
      '--model_dir',
      type=str,
      default='/tmp/imagenet',
      help="""\
      Path to classify_image_graph_def.pb,
      imagenet_synset_to_human_label_map.txt, and
      imagenet_2012_challenge_label_map_proto.pbtxt.\
      """
  )
  parser.add_argument(
      '--image_file',
      type=str,
      default='',
      help='Absolute path to image file.'
  )
  parser.add_argument(
      '--num_top_predictions',
      type=int,
      default=5,
      help='Display this many predictions.'
  )
  FLAGS, unparsed = parser.parse_known_args()

  tf.app.run(main=main, argv=[sys.argv[0]] + unparsed)

结果：

 runfile('F:/AI CODE/models-master/tutorials/image/imagenet/classify_image.py', wdir='F:/AI CODE/models-master/tutorials/image/imagenet')
giant panda, panda, panda bear, coon bear, Ailuropoda melanoleuca (score = 0.89632)
indri, indris, Indri indri, Indri brevicaudatus (score = 0.00766)
lesser panda, red panda, panda, bear cat, cat bear, Ailurus fulgens (score = 0.00266)
custard apple (score = 0.00138)
earthstar (score = 0.00104

http://blog.csdn.net/xiaopihaierletian/article/details/64507376