SingleClassRL

Implementation of weak single object segmentation from paper "Regularized Loss for Weakly Supervised Single Class Semantic Segmentation", ECCV2020. PDF

Main Files

train_with_anneal.py: use for training first in annealing stage, then in normal stage
train_with_transfer.py: use from training with weight transfer from another dataset, models that can be used for weight transfer are in directory 'trained_models'

OxfodPet dataset

Download OxfordPet from (https://www.robots.ox.ac.uk/~vgg/data/pets/) Files in 'SingleClassRL\data\Oxford_iit_pet\annotations' should be placed in the 'annotation' directory of OxfordPet dataset

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
data/Oxford_iit_pet/annotations		data/Oxford_iit_pet/annotations
trained_models		trained_models
.gitignore		.gitignore
README.md		README.md
custom_transforms.py		custom_transforms.py
evaluator.py		evaluator.py
get_dataset.py		get_dataset.py
get_losses.py		get_losses.py
get_model.py		get_model.py
get_standard_arguments.py		get_standard_arguments.py
image_utils.py		image_utils.py
loss_utils.py		loss_utils.py
losses.py		losses.py
other_utils.py		other_utils.py
oxford_iii_pet.py		oxford_iii_pet.py
param_functions.py		param_functions.py
train_epoch.py		train_epoch.py
train_with_anneal.py		train_with_anneal.py
train_with_transfer.py		train_with_transfer.py
training_utils.py		training_utils.py
unet_models.py		unet_models.py

morduspordus/SingleClassRL

Folders and files

Latest commit

History

Repository files navigation

SingleClassRL

Main Files

OxfodPet dataset

About

Resources

Stars

Watchers

Forks

Languages