Abstract
In this work, we introduce Scribbles for All, a label and training data
generation algorithm for semantic segmentation trained on scribble labels.
Training or fine-tuning semantic segmentation models with weak supervision has
become an important topic recently and was subject to significant advances in
model quality. In this setting, scribbles are a promising label type to achieve
high quality segmentation results while requiring a much lower annotation
effort than usual pixel-wise dense semantic segmentation annotations. The main
limitation of scribbles as source for weak supervision is the lack of
challenging datasets for scribble segmentation, which hinders the development
of novel methods and conclusive evaluations. To overcome this limitation,
Scribbles for All provides scribble labels for several popular segmentation
datasets and provides an algorithm to automatically generate scribble labels
for any dataset with dense annotations, paving the way for new insights and
model advancements in the field of weakly supervised segmentation. In addition
to providing datasets and algorithm, we evaluate state-of-the-art segmentation
models on our datasets and show that models trained with our synthetic labels
perform competitively with respect to models trained on manual labels. Thus,
our datasets enable state-of-the-art research into methods for scribble-labeled
semantic segmentation. The datasets, scribble generation algorithm, and
baselines are publicly available at github.com/wbkit/Scribbles4All
BibTeX
@online{Boettcher_2408.12489,
TITLE = {{S}cribbles for All: {B}enchmarking Scribble Supervised Segmentation Across Datasets},
AUTHOR = {B{\"o}ttcher, Wolfgang and Hoyer, Lukas and Unal, Ozan and Lenssen, Jan Eric and Schiele, Bernt},
LANGUAGE = {eng},
URL = {https://arxiv.org/abs/2408.12489},
EPRINT = {2408.12489},
EPRINTTYPE = {arXiv},
YEAR = {2024},
MARGINALMARK = {$\bullet$},
ABSTRACT = {In this work, we introduce Scribbles for All, a label and training data<br>generation algorithm for semantic segmentation trained on scribble labels.<br>Training or fine-tuning semantic segmentation models with weak supervision has<br>become an important topic recently and was subject to significant advances in<br>model quality. In this setting, scribbles are a promising label type to achieve<br>high quality segmentation results while requiring a much lower annotation<br>effort than usual pixel-wise dense semantic segmentation annotations. The main<br>limitation of scribbles as source for weak supervision is the lack of<br>challenging datasets for scribble segmentation, which hinders the development<br>of novel methods and conclusive evaluations. To overcome this limitation,<br>Scribbles for All provides scribble labels for several popular segmentation<br>datasets and provides an algorithm to automatically generate scribble labels<br>for any dataset with dense annotations, paving the way for new insights and<br>model advancements in the field of weakly supervised segmentation. In addition<br>to providing datasets and algorithm, we evaluate state-of-the-art segmentation<br>models on our datasets and show that models trained with our synthetic labels<br>perform competitively with respect to models trained on manual labels. Thus,<br>our datasets enable state-of-the-art research into methods for scribble-labeled<br>semantic segmentation. The datasets, scribble generation algorithm, and<br>baselines are publicly available at https://github.com/wbkit/Scribbles4All<br>},
}Endnote
%0 Report %A Böttcher, Wolfgang %A Hoyer, Lukas %A Unal, Ozan %A Lenssen, Jan Eric %A Schiele, Bernt %+ Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society External Organizations External Organizations Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society %T Scribbles for All: Benchmarking Scribble Supervised Segmentation Across Datasets : %G eng %U http://hdl.handle.net/21.11116/0000-0010-4548-0 %U https://arxiv.org/abs/2408.12489 %D 2024 %X In this work, we introduce Scribbles for All, a label and training data<br>generation algorithm for semantic segmentation trained on scribble labels.<br>Training or fine-tuning semantic segmentation models with weak supervision has<br>become an important topic recently and was subject to significant advances in<br>model quality. In this setting, scribbles are a promising label type to achieve<br>high quality segmentation results while requiring a much lower annotation<br>effort than usual pixel-wise dense semantic segmentation annotations. The main<br>limitation of scribbles as source for weak supervision is the lack of<br>challenging datasets for scribble segmentation, which hinders the development<br>of novel methods and conclusive evaluations. To overcome this limitation,<br>Scribbles for All provides scribble labels for several popular segmentation<br>datasets and provides an algorithm to automatically generate scribble labels<br>for any dataset with dense annotations, paving the way for new insights and<br>model advancements in the field of weakly supervised segmentation. In addition<br>to providing datasets and algorithm, we evaluate state-of-the-art segmentation<br>models on our datasets and show that models trained with our synthetic labels<br>perform competitively with respect to models trained on manual labels. Thus,<br>our datasets enable state-of-the-art research into methods for scribble-labeled<br>semantic segmentation. The datasets, scribble generation algorithm, and<br>baselines are publicly available at https://github.com/wbkit/Scribbles4All<br> %K Computer Science, Computer Vision and Pattern Recognition, cs.CV