8.1.1.3.1.4. blueoil.datasets.camvid

8.1.1.3.1.4.1. Module Contents

8.1.1.3.1.4.1.1. Classes

CamvidBase

Base class for CamVid and the variant dataset formats.

Camvid

CamVid

CamvidCustom

CamvidCustom

8.1.1.3.1.4.1.2. Functions

get_image(filename, convert_rgb=True, ignore_class_idx=None)

Returns numpy array of an image

blueoil.datasets.camvid.get_image(filename, convert_rgb=True, ignore_class_idx=None)

Returns numpy array of an image

class blueoil.datasets.camvid.CamvidBase(batch_size=10, *args, **kwargs)

Bases: blueoil.datasets.base.SegmentationBase

Base class for CamVid and the variant dataset formats.

http://www0.cs.ucl.ac.uk/staff/G.Brostow/papers/Brostow_2009-PRL.pdf https://github.com/alexgkendall/SegNet-Tutorial/tree/master/CamVid

extend_dir = CamVid
ignore_class_idx
property num_per_epoch(self)

Returns the number of datas in the data subset.

files_and_annotations(self)

Return all files and gt_boxes list.

__getitem__(self, i)

Returns the i-th item of the dataset.

__len__(self)

returns the number of items in the dataset.

class blueoil.datasets.camvid.Camvid(batch_size=10, *args, **kwargs)

Bases: blueoil.datasets.camvid.CamvidBase

CamVid

Original CamVid dataset format. http://www0.cs.ucl.ac.uk/staff/G.Brostow/papers/Brostow_2009-PRL.pdf https://github.com/alexgkendall/SegNet-Tutorial/tree/master/CamVid

IMAGE_HEIGHT = 360
IMAGE_WIDTH = 480
NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN = 367
NUM_EXAMPLES_PER_EPOCH_FOR_TEST = 101
classes = ['sky', 'building', 'pole', 'road', 'pavement', 'tree', 'signsymbol', 'fence', 'car', 'pedestrian', 'bicyclist']
num_classes
property label_colors(self)
property files_and_annotations(self)

Return all files and gt_boxes list.

class blueoil.datasets.camvid.CamvidCustom(batch_size=10, *args, **kwargs)

Bases: blueoil.datasets.base.StoragePathCustomizable, blueoil.datasets.camvid.CamvidBase

CamvidCustom

CamVid base custom dataset format.

property label_colors(self)
property classes(self)

Return the classes list in the data set.

property num_classes(self)

Return the number of classes in the data set.

parse_labels(self)
property files_and_annotations(self)

Return image and annotation file list. If there is no test dataset, then split dataset to train and test lists with specific ratio.