File Download

There are no files associated with this item.

Supplementary

Conference Paper: Visualizing the Invisible: Occluded Vehicle Segmentation and Recovery

TitleVisualizing the Invisible: Occluded Vehicle Segmentation and Recovery
Authors
Issue Date2019
PublisherComputer Vision Foundation. The Conference web site is located at http://openaccess.thecvf.com/ICCV2019.py
Citation
IEEE International Conference on Computer Vision (ICCV 2019), Seoul, Korea, 27 October - 2 November 2019, p. 7618-7627 How to Cite?
AbstractIn this paper, we propose a novel iterative multi-task framework to complete the segmentation mask of an occluded vehicle and recover the appearance of its invisible parts. In particular, to improve the quality of the segmentation completion, we present two coupled discriminators and introduce an auxiliary 3D model pool for sampling authentic silhouettes as adversarial samples. In addition, we propose a two-path structure with a shared network to enhance the appearance recovery capability. By iteratively performing the segmentation completion and the appearance recovery, the results will be progressively refined. To evaluate our method, we present a dataset, the Occluded Vehicle dataset, containing synthetic and real-world occluded vehicle images. We conduct comparison experiments on this dataset and demonstrate that our model outperforms the state-of-the-art in tasks of recovering segmentation mask and appearance for occluded vehicles. Moreover, we also demonstrate that our appearance recovery approach can benefit the occluded vehicle tracking in real-world videos.
Persistent Identifierhttp://hdl.handle.net/10722/273023

 

DC FieldValueLanguage
dc.contributor.authorYan, XS-
dc.contributor.authorYu, YL-
dc.contributor.authorWang, FGG-
dc.contributor.authorLiu, WX-
dc.contributor.authorHe, SF-
dc.contributor.authorPan, J-
dc.date.accessioned2019-08-06T09:21:05Z-
dc.date.available2019-08-06T09:21:05Z-
dc.date.issued2019-
dc.identifier.citationIEEE International Conference on Computer Vision (ICCV 2019), Seoul, Korea, 27 October - 2 November 2019, p. 7618-7627-
dc.identifier.urihttp://hdl.handle.net/10722/273023-
dc.description.abstractIn this paper, we propose a novel iterative multi-task framework to complete the segmentation mask of an occluded vehicle and recover the appearance of its invisible parts. In particular, to improve the quality of the segmentation completion, we present two coupled discriminators and introduce an auxiliary 3D model pool for sampling authentic silhouettes as adversarial samples. In addition, we propose a two-path structure with a shared network to enhance the appearance recovery capability. By iteratively performing the segmentation completion and the appearance recovery, the results will be progressively refined. To evaluate our method, we present a dataset, the Occluded Vehicle dataset, containing synthetic and real-world occluded vehicle images. We conduct comparison experiments on this dataset and demonstrate that our model outperforms the state-of-the-art in tasks of recovering segmentation mask and appearance for occluded vehicles. Moreover, we also demonstrate that our appearance recovery approach can benefit the occluded vehicle tracking in real-world videos.-
dc.languageeng-
dc.publisherComputer Vision Foundation. The Conference web site is located at http://openaccess.thecvf.com/ICCV2019.py-
dc.relation.ispartofInternational Conference on Computer Vision (ICCV)-
dc.titleVisualizing the Invisible: Occluded Vehicle Segmentation and Recovery-
dc.typeConference_Paper-
dc.identifier.emailPan, J: jpan@cs.hku.hk-
dc.identifier.authorityPan, J=rp01984-
dc.identifier.hkuros300346-
dc.identifier.spage7618-
dc.identifier.epage7627-
dc.publisher.placeUnited States-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats