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.…
Document Type: | Conference Proceeding |
---|---|
Conference Type: | Konferenzartikel |
Zitierlink: | https://opus.hs-offenburg.de/3954 | Bibliografische Angaben |
Title (English): | Object Segmentation Using Pixel-Wise Adversarial Loss |
Conference: | 41st DAGM German Conference (DAGM GCPR 2019), September 10-13, 2019, Dortmund, Germany |
Author: | Ricard Durall Lopez, Franz-Josef Pfreundt, Janis KeuperStaff MemberORCiDGND |
Year of Publication: | 2019 |
Place of publication: | Cham |
Publisher: | Springer |
First Page: | 303 |
Last Page: | 316 |
Parent Title (English): | DAGM GCPR 2019: Pattern Recognition |
Volume: | LNCS 11824 |
ISBN: | 978-3-030-33676-9 (Online) |
ISBN: | 978-3-030-33675-2 (Print) |
ISSN: | 1611-3349 (Online) |
ISSN: | 0302-9743 (Print) |
DOI: | https://doi.org/10.1007/978-3-030-33676-9_21 |
Language: | English | Inhaltliche Informationen |
Institutes: | Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019) |
Institutes: | Bibliografie |
DDC classes: | 000 Allgemeines, Informatik, Informationswissenschaft | Formale Angaben |
Open Access: | Closed Access |
Licence (German): | Urheberrechtlich geschützt |