# coding=utf-8 # Copyright 2020 The TensorFlow Datasets Authors. # # 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. # Lint as: python3 """PASCAL VOC datasets.""" from __future__ import absolute_import from __future__ import division from __future__ import print_function import os import xml.etree.ElementTree import tensorflow.compat.v2 as tf import tensorflow_datasets.public_api as tfds _VOC_CITATION = """\ @misc{{pascal-voc-{year}, author = "Everingham, M. and Van~Gool, L. and Williams, C. K. I. and Winn, J. and Zisserman, A.", title = "The {{PASCAL}} {{V}}isual {{O}}bject {{C}}lasses {{C}}hallenge {year} {{(VOC{year})}} {{R}}esults", howpublished = "http://www.pascal-network.org/challenges/VOC/voc{year}/workshop/index.html"}} """ _VOC_DESCRIPTION = """\ This dataset contains the data from the PASCAL Visual Object Classes Challenge {year}, a.k.a. VOC{year}, corresponding to the Classification and Detection competitions. A total of {num_images} images are included in this dataset, where each image contains a set of objects, out of 20 different classes, making a total of {num_objects} annotated objects. In the Classification competition, the goal is to predict the set of labels contained in the image, while in the Detection competition the goal is to predict the bounding box and label of each individual object. WARNING: As per the official dataset, the test set of VOC2012 does not contain annotations. """ _VOC_URL = "http://host.robots.ox.ac.uk/pascal/VOC/voc{year}/" # Original site, it is down very often. # _VOC_DATA_URL = "http://host.robots.ox.ac.uk/pascal/VOC/voc{year}/" # Data mirror: _VOC_DATA_URL = "http://pjreddie.com/media/files/" _VOC_LABELS = ( "aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor", ) _VOC_POSES = ( "frontal", "rear", "left", "right", "unspecified", ) def _get_example_objects(annon_filepath): """Function to get all the objects from the annotation XML file.""" with tf.io.gfile.GFile(annon_filepath, "r") as f: root = xml.etree.ElementTree.parse(f).getroot() # Disable pytype to avoid attribute-error due to find returning # Optional[Element] # pytype: disable=attribute-error size = root.find("size") width = float(size.find("width").text) height = float(size.find("height").text) for obj in root.findall("object"): # Get object's label name. label = obj.find("name").text.lower() # Get objects' pose name. pose = obj.find("pose").text.lower() is_truncated = (obj.find("truncated").text == "1") is_difficult = (obj.find("difficult").text == "1") bndbox = obj.find("bndbox") xmax = float(bndbox.find("xmax").text) xmin = float(bndbox.find("xmin").text) ymax = float(bndbox.find("ymax").text) ymin = float(bndbox.find("ymin").text) yield { "label": label, "pose": pose, "bbox": tfds.features.BBox( ymin / height, xmin / width, ymax / height, xmax / width), "is_truncated": is_truncated, "is_difficult": is_difficult, } # pytype: enable=attribute-error class VocConfig(tfds.core.BuilderConfig): """BuilderConfig for Voc.""" def __init__( self, year=None, filenames=None, has_test_annotations=True, **kwargs): self.year = year self.filenames = filenames self.has_test_annotations = has_test_annotations super(VocConfig, self).__init__( name=year, # Version history: # 4.0.0: Added BuildConfig and 2012 version support, deprecate Voc2007. # 3.0.0: S3 with new hashing function (different shuffle). # 2.0.0: S3 (new shuffling, sharding and slicing mechanism). version=tfds.core.Version("4.0.0"), **kwargs) class Voc(tfds.core.GeneratorBasedBuilder): """Pascal VOC 2007 or 2012.""" BUILDER_CONFIGS = [ VocConfig( year="2007", description=_VOC_DESCRIPTION.format( year=2007, num_images=9963, num_objects=24640), filenames={ "trainval": "VOCtrainval_06-Nov-2007.tar", "test": "VOCtest_06-Nov-2007.tar", }, has_test_annotations=True, ), VocConfig( year="2012", description=_VOC_DESCRIPTION.format( year=2012, num_images=11540, num_objects=27450), filenames={ "trainval": "VOCtrainval_11-May-2012.tar", "test": "VOC2012test.tar", }, has_test_annotations=False, ), ] def _info(self): return tfds.core.DatasetInfo( builder=self, description=self.builder_config.description, features=tfds.features.FeaturesDict({ "image": tfds.features.Image(), "image/filename": tfds.features.Text(), "objects": tfds.features.Sequence({ "label": tfds.features.ClassLabel(names=_VOC_LABELS), "bbox": tfds.features.BBoxFeature(), "pose": tfds.features.ClassLabel(names=_VOC_POSES), "is_truncated": tf.bool, "is_difficult": tf.bool, }), "labels": tfds.features.Sequence( tfds.features.ClassLabel(names=_VOC_LABELS)), "labels_no_difficult": tfds.features.Sequence( tfds.features.ClassLabel(names=_VOC_LABELS)), }), homepage=_VOC_URL.format(year=self.builder_config.year), citation=_VOC_CITATION.format(year=self.builder_config.year), ) def _split_generators(self, dl_manager): paths = dl_manager.download_and_extract({ k: os.path.join(_VOC_DATA_URL, v) for k, v in self.builder_config.filenames.items() }) return [ tfds.core.SplitGenerator( name=tfds.Split.TEST, gen_kwargs=dict(data_path=paths["test"], set_name="test")), tfds.core.SplitGenerator( name=tfds.Split.TRAIN, gen_kwargs=dict(data_path=paths["trainval"], set_name="train")), tfds.core.SplitGenerator( name=tfds.Split.VALIDATION, gen_kwargs=dict(data_path=paths["trainval"], set_name="val")), ] def _generate_examples(self, data_path, set_name): """Yields examples.""" set_filepath = os.path.join( data_path, os.path.normpath("VOCdevkit/VOC{}/ImageSets/Main/{}.txt".format( self.builder_config.year, set_name))) load_annotations = ( self.builder_config.has_test_annotations or set_name != "test") with tf.io.gfile.GFile(set_filepath, "r") as f: for line in f: image_id = line.strip() example = self._generate_example(data_path, image_id, load_annotations) yield image_id, example def _generate_example(self, data_path, image_id, load_annotations): image_filepath = os.path.join( data_path, os.path.normpath("VOCdevkit/VOC{}/JPEGImages/{}.jpg".format( self.builder_config.year, image_id))) annon_filepath = os.path.join( data_path, os.path.normpath("VOCdevkit/VOC{}/Annotations/{}.xml".format( self.builder_config.year, image_id))) if load_annotations: objects = list(_get_example_objects(annon_filepath)) # Use set() to remove duplicates labels = sorted(set(obj["label"] for obj in objects)) labels_no_difficult = sorted(set( obj["label"] for obj in objects if obj["is_difficult"] == 0 )) else: # The test set of VOC2012 does not contain annotations objects = [] labels = [] labels_no_difficult = [] return { "image": image_filepath, "image/filename": image_id + ".jpg", "objects": objects, "labels": labels, "labels_no_difficult": labels_no_difficult, }