# 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.
"""Defines KITTI segmentation, including the MSeg version.
KITTI has a wide setup of dataset configurations aimed for benchmarking
different modalities (flow, depth, segmentation, detection, ...), see:
http://www.cvlibs.net/datasets/kitti/
Paper:
Vision meets Robotics: The KITTI Dataset.
Andreas Geiger and Philip Lenz and Christoph Stiller and Raquel Urtasun.
International Journal of Robotics Research (IJRR), 2013
"""
from factors_of_influence.fids import mseg_base
KITTISeg = mseg_base.MSegBase(
mseg_name='KITTI Segmentation',
mseg_original_name='kitti-34',
mseg_base_name='kitti-19',
mseg_dirname='KITTI/',
mseg_train_dataset=False,
mseg_segmentation_background_labels=['unlabeled', 'out of roi'],
)
Pascal Context:
URL: https://www.cs.stanford.edu/~roozbeh/pascal-context/
Paper:
- The Role of Context for Object Detection and Semantic Segmentation in the Wild
Roozbeh Mottaghi, Xianjie Chen, Xiaobai Liu, Nam-Gyu Cho, Seong-Whan Lee,
Sanja Fidler, Raquel Urtasun, Alan Yuille. In CVPR, 2014.
"""
from factors_of_influence.fids import mseg_base
PascalContext = mseg_base.MSegBase(
mseg_name='Pascal Context',
mseg_original_name='pascal-context-460',
mseg_base_name='pascal-context-60',
mseg_dirname='PASCAL_Context/',
mseg_train_dataset=False,
mseg_segmentation_background_labels=[
'background', 'unlabeled', 'Unlabeled'
],
mseg_use_mapping_for_mseg_segmentation=True,
)
PascalVOC = mseg_base.MSegBase(
mseg_name='Pascal VOC2012',
mseg_original_name='voc2012',
mseg_base_name='voc2012',
mseg_dirname='PASCAL_VOC_2012/',
mseg_train_dataset=False,
mseg_segmentation_background_labels=['background'],
mseg_use_mapping_for_mseg_segmentation=True,
)
# 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.
"""Defines MapillaryVistas Public dataset, including the MSeg version.
URL: https://www.mapillary.com/dataset/vistas
Paper:
The Mapillary Vistas Dataset for semantic understanding of street scenes.
G. Neuhold, T. Ollmann, S. Rota Bulo, and P. Kontschieder. In ICCV, 2017.
"""
from factors_of_influence.fids import mseg_base
MapillaryVistasPublic = mseg_base.MSegBase(
mseg_name='Mapillary Vistas Public (MVD)',
mseg_original_name='mapillary-public66',
mseg_base_name='mapillary-public65',
mseg_dirname='MapillaryVistasPublic/',
mseg_train_dataset=True,
)
"""Defines IDD (Indian Driving Dataset), including the MSeg version.
URL: https://idd.insaan.iiit.ac.in/dataset/details/
Paper:
IDD: A Dataset for Exploring Problems of Autonomous Navigation
in Unconstrained Environments. Girish Varma, Anbumani Subramanian,
Anoop Namboodiri, Manmohan Chandraker, and C V Jawahar. In WACV, 2019.
Note: The detection dataset (also available for download) has roughly 4 times
more annotated images available than the segmentation dataset.
"""
from factors_of_influence.fids import mseg_base
IDD = mseg_base.MSegBase(
mseg_name='Indian Driving Dataset (IDD)',
mseg_original_name='idd-40',
mseg_base_name='idd-39',
mseg_dirname='IDD/IDD_Segmentation/',
mseg_train_dataset=True,
mseg_segmentation_background_labels=[
'unlabeled', 'out of roi', 'train', 'ego vehicle',
'rectification border', 'license plate'
], # These classes are either not present in the official label set, but
# yet are present in the annotation, or are annotated as not for training.
# See: https://idd.insaan.iiit.ac.in/dataset/details/
mseg_use_mapping_for_mseg_segmentation=True,
)
# 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.
"""Defines WildDash-V1, including the MSeg version.
URL: https://wilddash.cc/
Paper:
Wilddash - creating hazard-aware benchmarks.
O. Zendel, K. Honauer, M. Murschitz, D. Steininger, and G. Fernandez Dominguez.
In ECCV, 2018.
"""
from factors_of_influence.fids import mseg_base
WildDash19 = mseg_base.MSegBase(
mseg_name='WildDashDataset',
mseg_original_name='wilddash-34',
mseg_base_name='wilddash-19',
mseg_dirname='WildDash/',
mseg_train_dataset=False,
mseg_segmentation_background_labels=['unlabeled', 'out of roi'],
)
# 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.
"""Defines ScanNet, including the MSeg version.
URL: http://www.scan-net.org/
Paper:
ScanNet: Richly-annotated 3d reconstructions of indoor scenes.
A. Dai, A. X. Chang, M. Savva, M. Halber, T. Funkhouser, and M. Nießner.
In CVPR, 2017.
"""
from factors_of_influence.fids import mseg_base
ScanNet20 = mseg_base.MSegBase(
mseg_name='ScanNet',
mseg_original_name='scannet-41',
mseg_base_name='scannet-20',
mseg_dirname='ScanNet/scannet_frames_25k/',
mseg_train_dataset=False,
)
# 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.
"""Imports CityScapes, only the MSeg version.
URL: https://www.cityscapes-dataset.com/downloads/
Paper:
The cityscapes dataset for semantic urban scene understanding.
M. Cordts, M. Omran, S. Ramos, T. Rehfeld, M. Enzweiler, R. Benenson,
U. Franke, S. Roth, and B. Schiele. In CVPR, 2016.
City Scapes contains additional modalities / ground-truths, including:
- disparity (depth)
- coarse segmentation
- person boundingboxes.
"""
from factors_of_influence.fids import mseg_base
CityScapes = mseg_base.MSegBase(
mseg_name='CityScapes',
mseg_original_name='cityscapes-34',
mseg_base_name='cityscapes-19',
mseg_dirname='Cityscapes/',
mseg_train_dataset=True,
mseg_segmentation_background_labels=['unlabeled', 'out of roi'],
)