create_pascal_tfrecord.py

# Copyright 2020 Google Research. 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.
# ==============================================================================
r"""Convert PASCAL dataset to TFRecord.

Example usage:
    python create_pascal_tfrecord.py  --data_dir=/tmp/VOCdevkit  \
        --year=VOC2012  --output_path=/tmp/pascal
"""
import hashlib
import io
import json
import os

from absl import app
from absl import flags
from absl import logging

from lxml import etree
import PIL.Image
import tensorflow as tf

from dataset import tfrecord_util

FLAGS = flags.FLAGS

SETS = ['train', 'val', 'trainval', 'test']
YEARS = ['VOC2007', 'VOC2012', 'merged']

pascal_label_map_dict = {
    'background': 0,
    'aeroplane': 1,
    'bicycle': 2,
    'bird': 3,
    'boat': 4,
    'bottle': 5,
    'bus': 6,
    'car': 7,
    'cat': 8,
    'chair': 9,
    'cow': 10,
    'diningtable': 11,
    'dog': 12,
    'horse': 13,
    'motorbike': 14,
    'person': 15,
    'pottedplant': 16,
    'sheep': 17,
    'sofa': 18,
    'train': 19,
    'tvmonitor': 20,
}

GLOBAL_IMG_ID = 0  # global image id.
GLOBAL_ANN_ID = 0  # global annotation id.


def define_flags():
  """Define the flags."""
  flags.DEFINE_string('data_dir', '',
                      'Root directory to raw PASCAL VOC dataset.')
  flags.DEFINE_string('set', 'train', 'Convert training set, validation set or '
                      'merged set.')
  flags.DEFINE_string('annotations_dir', 'Annotations',
                      '(Relative) path to annotations directory.')
  flags.DEFINE_string('year', 'VOC2007', 'Desired challenge year.')
  flags.DEFINE_string('output_path', '', 'Path to output TFRecord and json.')
  flags.DEFINE_string('label_map_json_path', None,
                      'Path to label map json file with a dictionary.')
  flags.DEFINE_boolean('ignore_difficult_instances', False, 'Whether to ignore '
                       'difficult instances')
  flags.DEFINE_integer('num_shards', 100, 'Number of shards for output file.')
  flags.DEFINE_integer('num_images', None, 'Max number of imags to process.')


def get_image_id(filename):
  """Convert a string to a integer."""
  # Warning: this function is highly specific to pascal filename!!
  # Given filename like '2008_000002', we cannot use id 2008000002 because our
  # code internally will convert the int value to float32 and back to int, which
  # would cause value mismatch int(float32(2008000002)) != int(2008000002).
  # COCO needs int values, here we just use a incremental global_id, but
  # users should customize their own ways to generate filename.
  del filename
  global GLOBAL_IMG_ID
  GLOBAL_IMG_ID += 1
  return GLOBAL_IMG_ID


def get_ann_id():
  """Return unique annotation id across images."""
  global GLOBAL_ANN_ID
  GLOBAL_ANN_ID += 1
  return GLOBAL_ANN_ID


def dict_to_tf_example(data,
                       images_dir,
                       label_map_dict,
                       ignore_difficult_instances=False,
                       ann_json_dict=None):
  """Convert XML derived dict to tf.Example proto.

  Notice that this function normalizes the bounding box coordinates provided
  by the raw data.

  Args:
    data: dict holding PASCAL XML fields for a single image (obtained by running
      tfrecord_util.recursive_parse_xml_to_dict)
    images_dir: Path to the directory holding raw images.
    label_map_dict: A map from string label names to integers ids.
    ignore_difficult_instances: Whether to skip difficult instances in the
      dataset  (default: False).
    ann_json_dict: annotation json dictionary.

  Returns:
    example: The converted tf.Example.

  Raises:
    ValueError: if the image pointed to by data['filename'] is not a valid JPEG
  """
  full_path = os.path.join(images_dir, data['filename'])
  with tf.io.gfile.GFile(full_path, 'rb') as fid:
    encoded_jpg = fid.read()
  encoded_jpg_io = io.BytesIO(encoded_jpg)
  image = PIL.Image.open(encoded_jpg_io)
  if image.format != 'JPEG':
    raise ValueError('Image format not JPEG')
  key = hashlib.sha256(encoded_jpg).hexdigest()

  width = int(data['size']['width'])
  height = int(data['size']['height'])
  image_id = get_image_id(data['filename'])
  if ann_json_dict:
    image = {
        'file_name': data['filename'],
        'height': height,
        'width': width,
        'id': image_id,
    }
    ann_json_dict['images'].append(image)

  xmin = []
  ymin = []
  xmax = []
  ymax = []
  area = []
  classes = []
  classes_text = []
  truncated = []
  poses = []
  difficult_obj = []
  if 'object' in data:
    for obj in data['object']:
      difficult = bool(int(obj['difficult']))
      if ignore_difficult_instances and difficult:
        continue

      difficult_obj.append(int(difficult))

      xmin.append(float(obj['bndbox']['xmin']) / width)
      ymin.append(float(obj['bndbox']['ymin']) / height)
      xmax.append(float(obj['bndbox']['xmax']) / width)
      ymax.append(float(obj['bndbox']['ymax']) / height)
      area.append((xmax[-1] - xmin[-1]) * (ymax[-1] - ymin[-1]))
      classes_text.append(obj['name'].encode('utf8'))
      classes.append(label_map_dict[obj['name']])
      truncated.append(int(obj['truncated']))
      poses.append(obj['pose'].encode('utf8'))

      if ann_json_dict:
        abs_xmin = int(obj['bndbox']['xmin'])
        abs_ymin = int(obj['bndbox']['ymin'])
        abs_xmax = int(obj['bndbox']['xmax'])
        abs_ymax = int(obj['bndbox']['ymax'])
        abs_width = abs_xmax - abs_xmin
        abs_height = abs_ymax - abs_ymin
        ann = {
            'area': abs_width * abs_height,
            'iscrowd': 0,
            'image_id': image_id,
            'bbox': [abs_xmin, abs_ymin, abs_width, abs_height],
            'category_id': label_map_dict[obj['name']],
            'id': get_ann_id(),
            'ignore': 0,
            'segmentation': [],
        }
        ann_json_dict['annotations'].append(ann)

  example = tf.train.Example(
      features=tf.train.Features(
          feature={
              'image/height':
                  tfrecord_util.int64_feature(height),
              'image/width':
                  tfrecord_util.int64_feature(width),
              'image/filename':
                  tfrecord_util.bytes_feature(data['filename'].encode('utf8')),
              'image/source_id':
                  tfrecord_util.bytes_feature(str(image_id).encode('utf8')),
              'image/key/sha256':
                  tfrecord_util.bytes_feature(key.encode('utf8')),
              'image/encoded':
                  tfrecord_util.bytes_feature(encoded_jpg),
              'image/format':
                  tfrecord_util.bytes_feature('jpeg'.encode('utf8')),
              'image/object/bbox/xmin':
                  tfrecord_util.float_list_feature(xmin),
              'image/object/bbox/xmax':
                  tfrecord_util.float_list_feature(xmax),
              'image/object/bbox/ymin':
                  tfrecord_util.float_list_feature(ymin),
              'image/object/bbox/ymax':
                  tfrecord_util.float_list_feature(ymax),
              'image/object/area':
                  tfrecord_util.float_list_feature(area),
              'image/object/class/text':
                  tfrecord_util.bytes_list_feature(classes_text),
              'image/object/class/label':
                  tfrecord_util.int64_list_feature(classes),
              'image/object/difficult':
                  tfrecord_util.int64_list_feature(difficult_obj),
              'image/object/truncated':
                  tfrecord_util.int64_list_feature(truncated),
              'image/object/view':
                  tfrecord_util.bytes_list_feature(poses),
          }))
  return example


def main(_):
  if FLAGS.set not in SETS:
    raise ValueError('set must be in : {}'.format(SETS))
  if FLAGS.year not in YEARS:
    raise ValueError('year must be in : {}'.format(YEARS))
  if not FLAGS.output_path:
    raise ValueError('output_path cannot be empty.')

  data_dir = FLAGS.data_dir
  years = ['VOC2007', 'VOC2012']
  if FLAGS.year != 'merged':
    years = [FLAGS.year]

  output_dir = os.path.dirname(FLAGS.output_path)
  if not tf.io.gfile.exists(output_dir):
    tf.io.gfile.makedirs(output_dir)
  logging.info('Writing to output directory: %s', output_dir)

  writers = [
      tf.io.TFRecordWriter(FLAGS.output_path + '-%05d-of-%05d.tfrecord' %
                           (i, FLAGS.num_shards))
      for i in range(FLAGS.num_shards)
  ]

  if FLAGS.label_map_json_path:
    with tf.io.gfile.GFile(FLAGS.label_map_json_path, 'rb') as f:
      label_map_dict = json.load(f)
  else:
    label_map_dict = pascal_label_map_dict

  ann_json_dict = {
      'images': [],
      'type': 'instances',
      'annotations': [],
      'categories': []
  }
  for year in years:
    example_class = list(label_map_dict.keys())[1]
    examples_path = os.path.join(data_dir, year, 'ImageSets', 'Main',
                                 example_class + '_' + FLAGS.set + '.txt')
    examples_list = tfrecord_util.read_examples_list(examples_path)
    annotations_dir = os.path.join(data_dir, year, FLAGS.annotations_dir)

    for class_name, class_id in label_map_dict.items():
      cls = {'supercategory': 'none', 'id': class_id, 'name': class_name}
      ann_json_dict['categories'].append(cls)

    logging.info('Reading from PASCAL %s dataset.', year)
    for idx, example in enumerate(examples_list):
      if FLAGS.num_images and idx >= FLAGS.num_images:
        break
      if idx % 100 == 0:
        logging.info('On image %d of %d', idx, len(examples_list))
      path = os.path.join(annotations_dir, example + '.xml')
      with tf.io.gfile.GFile(path, 'r') as fid:
        xml_str = fid.read()
      xml = etree.fromstring(xml_str)
      data = tfrecord_util.recursive_parse_xml_to_dict(xml)['annotation']

      img_dir = os.path.join(FLAGS.data_dir, data['folder'], 'JPEGImages')
      tf_example = dict_to_tf_example(
          data,
          img_dir,
          label_map_dict,
          FLAGS.ignore_difficult_instances,
          ann_json_dict=ann_json_dict)
      writers[idx % FLAGS.num_shards].write(tf_example.SerializeToString())

  for writer in writers:
    writer.close()

  json_file_path = os.path.join(
      os.path.dirname(FLAGS.output_path),
      'json_' + os.path.basename(FLAGS.output_path) + '.json')
  with tf.io.gfile.GFile(json_file_path, 'w') as f:
    json.dump(ann_json_dict, f)


if __name__ == '__main__':
  define_flags()
  app.run(main)