Volltext-Downloads (blau) und Frontdoor-Views (grau)
  • search hit 27 of 176
Back to Result List

Object Segmentation Using Pixel-Wise Adversarial Loss

  • Recent deep learning based approaches have shown remarkable success on object segmentation tasks. However, there is still room for further improvement. Inspired by generative adversarial networks, we present a generic end-to-end adversarial approach, which can be combined with a wide range of existing semantic segmentation networks to improve their segmentation performance. The key element of ourRecent deep learning based approaches have shown remarkable success on object segmentation tasks. However, there is still room for further improvement. Inspired by generative adversarial networks, we present a generic end-to-end adversarial approach, which can be combined with a wide range of existing semantic segmentation networks to improve their segmentation performance. The key element of our method is to replace the commonly used binary adversarial loss with a high resolution pixel-wise loss. In addition, we train our generator employing stochastic weight averaging fashion, which further enhances the predicted output label maps leading to state-of-the-art results. We show, that this combination of pixel-wise adversarial training and weight averaging leads to significant and consistent gains in segmentation performance, compared to the baseline models.show moreshow less

Export metadata

Additional Services

Share in Twitter Search Google Scholar

Statistics

frontdoor_oas
Metadaten
Author:Ricard Durall Lopez, Franz-Josef Pfreundt, Janis KeuperORCiDGND
Publisher:Springer
Place of publication:Cham
Year of Publication:2019
ISBN:978-3-030-33676-9 (Online)
ISBN:978-3-030-33675-2 (Print)
Language:English
DDC classes:000 Allgemeines, Informatik, Informationswissenschaft
Parent Title (English):DAGM GCPR 2019: Pattern Recognition
Volume:LNCS 11824
ISSN:1611-3349 (Online)
ISSN:0302-9743 (Print)
First Page:303
Last Page:316
Document Type:Conference Proceeding
Institutes:Bibliografie
Open Access:Zugriffsbeschränkt
Release Date:2020/01/21
Licence (German):License LogoEs gilt das UrhG
Note:
41st DAGM German Conference, DAGM GCPR 2019, Dortmund, Germany, September 10–13, 2019, Proceedings
DOI:https://doi.org/10.1007/978-3-030-33676-9_21