def main():
LIST_OF_OBJECTS = []
count = 1
pbtxtFilepath = Path(PBTXT_FILE)
#if pbtxt file does not exist then create the file, or if it exists then remove it
if not pbtxtFilepath.exists():
open(PBTXT_FILE, 'w').close()
else:
os.system('rm ' + PBTXT_FILE)
for xmlFile in os.listdir(ANNOTATION_FILE_PATH):
#rotating through all the annotation files
with tf.gfile.GFile(ANNOTATION_FILE_PATH + '/' + xmlFile, 'r') as fid:
xml_str = fid.read()
xml = etree.fromstring(xml_str)
data = dataset_util.recursive_parse_xml_to_dict(xml)[
'annotation'] #reading from annotation tag
objects = data['object'] #reading the attributes of the object
filename = data['path'] #reading the path of the image file
for object in objects:
objectname = object['name'].strip() #getting the object name
if LIST_OF_OBJECTS.__contains__(objectname):
#if a particular object has already occured then move to the next object
continue
else:
#add the object to the object list
LIST_OF_OBJECTS.append(objectname)
writingToPbtxt(objectname) #write the object to the pbtxt file
count = count + 1
print('The Number of Classes in the Dataset is : ', len(LIST_OF_OBJECTS))
def create_tf_record(output_filename,
label_map_dict,
annotations_dir,
image_dir,
examples,
faces_only=True,
mask_type='png'):
"""Creates a TFRecord file from examples.
Args:
output_filename: Path to where output file is saved.
label_map_dict: The label map dictionary.
annotations_dir: Directory where annotation files are stored.
image_dir: Directory where image files are stored.
examples: Examples to parse and save to tf record.
faces_only: If True, generates bounding boxes for pet faces. Otherwise
generates bounding boxes (as well as segmentations for full pet bodies).
mask_type: 'numerical' or 'png'. 'png' is recommended because it leads to
smaller file sizes.
"""
writer = tf.python_io.TFRecordWriter(output_filename)
for idx, example in enumerate(examples):
if idx % 100 == 0:
logging.info('On image %d of %d', idx, len(examples))
xml_path = os.path.join(annotations_dir, 'xmls', example + '.xml')
mask_path = os.path.join(annotations_dir, 'trimaps', example + '.png')
if not os.path.exists(xml_path):
logging.warning('Could not find %s, ignoring example.', xml_path)
continue
with tf.gfile.GFile(xml_path, 'r') as fid:
xml_str = fid.read()
xml = etree.fromstring(xml_str)
data = dataset_util.recursive_parse_xml_to_dict(xml)['annotation']
try:
tf_example = dict_to_tf_example(data,
mask_path,
label_map_dict,
image_dir,
faces_only=faces_only,
mask_type=mask_type)
writer.write(tf_example.SerializeToString())
except ValueError:
logging.warning('Invalid example: %s, ignoring.', xml_path)
writer.close()
def create_tf_record(writer, label_map_dict, annotations_dir, image_dir,
examples, value):
#This function will create the TFR Record for both training and evaluation and will write it to the file
for idx, example in enumerate(examples):
if idx % 100 == 0:
logging.info('On image %d of %d', idx, len(examples))
path = os.path.join(annotations_dir, example.split('.')[0] + '.xml')
if not os.path.exists(path):
logging.warning('Could not find %s, ignoring example.', path)
continue
with tf.gfile.GFile(path, 'r') as fid:
xml_str = fid.read()
xml = etree.fromstring(xml_str)
data = dataset_util.recursive_parse_xml_to_dict(xml)['annotation']
tf_example = dict_to_tf_example(data, label_map_dict, image_dir, value)
writer.write(tf_example.SerializeToString())
def main():
#contains the class names from pbtxt file
label_map_dict = label_map_util.get_label_map_dict(FLAGS.label_map_path)
for l in label_map_dict:
print("label : " + l)
stringToWrite = ''
file_list = [] #contains all the files from the annotation file
c = 0
#taking all the files from the annotation directory
for xmlFile in os.listdir(ANNOTATION_FILE_PATH):
with tf.gfile.GFile(ANNOTATION_FILE_PATH + '/' + xmlFile,
'r') as fid:
xml_str = fid.read()
xml = etree.fromstring(xml_str)
data = dataset_util.recursive_parse_xml_to_dict(xml)[
'annotation'] #Getting data from annotation tag
imageFile = data['filename'] #Getting data for filename
objects = data[
'object'] #Getting datails of an object from xml file
objectname = ''
allObjects = []
for object in objects:
objectname = object['name'].strip()
allObjects.append(objectname)
if not allObjects.__contains__(l):
continue
else:
#Adding the files from xml files into a list
if not file_list.__contains__(imageFile):
file_list.append(imageFile)
LIST_OF_IMAGE_FILES = getListOfImgFiles()
stringToWrite = creatingTrainingFile(LIST_OF_IMAGE_FILES, file_list,
stringToWrite)
TRAINING_FILE = TRAINING_FILE_PATH + '/' + l + '_train.txt'
trainingfileopen = open(TRAINING_FILE, 'w')
trainingfileopen.write(stringToWrite)
trainingfileopen.close()
print('created')
def load_pascal_voc(origin_dir):
data = {}
cwd = os.getcwd()
try:
os.chdir(origin_dir)
except OSError:
print('Could not change directory.')
for file in glob.glob("*.xml"):
with tf.io.gfile.GFile(file, 'r') as fid:
xml_str = fid.read()
xml = etree.fromstring(xml_str)
data[file] = dataset_util.recursive_parse_xml_to_dict(
xml)['annotation']
try:
os.chdir(cwd)
except OSError:
print('Could not return to previous directory.')
return data
def create_tf_record(output_filename,
num_shards,
label_map_dict,
annotations_dir,
examples,
image_dir=FLAGS.data_dir,
faces_only=False):
"""Creates a TFRecord file from examples.
Args:
output_filename: Path to where output file is saved.
num_shards: Number of shards for output file.
label_map_dict: The label map dictionary.
annotations_dir: Directory where annotation files are stored.
image_dir: Directory where image files are stored.
examples: Examples to parse and save to tf record.
faces_only: If True, generates bounding boxes for pet faces. Otherwise
generates bounding boxes (as well as segmentations for full pet bodies).
mask_type: 'numerical' or 'png'. 'png' is recommended because it leads to
smaller file sizes.
"""
with contextlib2.ExitStack() as tf_record_close_stack:
output_tfrecords = tf_record_creation_util.open_sharded_output_tfrecords(
tf_record_close_stack, output_filename, num_shards)
for idx, example in enumerate(examples):
if idx % 100 == 0:
logging.info('On image %d of %d', idx, len(examples))
path = os.path.join(annotations_dir, example + '.xml')
with tf.gfile.GFile(path, 'r') as fid:
xml_str = fid.read()
xml = etree.fromstring(xml_str)
data = dataset_util.recursive_parse_xml_to_dict(xml)['annotation']
try:
tf_example = dict_to_tf_example(data, FLAGS.data_dir, label_map_dict,
FLAGS.ignore_difficult_instances)
if tf_example:
shard_idx = idx % num_shards
output_tfrecords[shard_idx].write(tf_example.SerializeToString())
except ValueError:
logging.warning('Invalid example: %s, ignoring.', path)
import sys
progPath = sys.argv[1]
Config = configparser.ConfigParser()
Config._interpolation = configparser.ExtendedInterpolation()
Config.read(progPath + "/config.ini")
IMG_TO_OBJECT_FILE_PATH = Config.get('tensorflow', 'IMG_TO_OBJECT_FILE_PATH')
fileopen = open(IMG_TO_OBJECT_FILE_PATH + '/imgnameToclassname.txt', 'w')
ANNOTATION_FILE_PATH = Config.get('tensorflow', 'ANNOTATION_FILE_PATH')
for xmlFile in os.listdir(ANNOTATION_FILE_PATH):
print('xml file : ' + xmlFile)
with tf.gfile.GFile(ANNOTATION_FILE_PATH + '/' + xmlFile, 'r') as fid:
xml_str = fid.read()
xml = etree.fromstring(xml_str)
data = dataset_util.recursive_parse_xml_to_dict(xml)['annotation']
objects = data['object']
objectname = ""
obj = []
for object in objects:
if obj.__contains__(object['name']):
continue
obj.append(object['name'])
objectname = objectname + " " + object['name']
imageFile = data['filename']
fileopen.write(imageFile + '=' + objectname + '\n')
fileopen.close()