Sökning: onr:"swepub:oai:DiVA.org:kth-315688" >
Scaling Semantic Se...
Scaling Semantic Segmentation Beyond 1K Classes on a Single GPU
-
- Jain, Shipra (författare)
- KTH,Skolan för elektroteknik och datavetenskap (EECS),Swiss Fed Inst Technol, Comp Vis Lab, Zurich, Switzerland.
-
- Paudel, Danda Pani (författare)
- Swiss Fed Inst Technol, Comp Vis Lab, Zurich, Switzerland.
-
- Danelljan, Martin (författare)
- Swiss Fed Inst Technol, Comp Vis Lab, Zurich, Switzerland.
-
visa fler...
-
- Van Gool, Luc (författare)
- Swiss Fed Inst Technol, Comp Vis Lab, Zurich, Switzerland.;Katholieke Univ Leuven, Leuven, Belgium.
-
visa färre...
-
(creator_code:org_t)
- Institute of Electrical and Electronics Engineers (IEEE), 2021
- 2021
- Engelska.
-
Ingår i: Proceedings of the IEEE International Conference on Computer Vision. - : Institute of Electrical and Electronics Engineers (IEEE). ; , s. 7406-7416
- Relaterad länk:
-
https://iccv2021.the...
-
visa fler...
-
https://urn.kb.se/re...
-
https://doi.org/10.1...
-
visa färre...
Abstract
Ämnesord
Stäng
- The state-of-the-art object detection and image classification methods can perform impressively on more than 9k classes. In contrast, the number of classes in semantic segmentation datasets is relatively limited. This is not surprising when the restrictions caused by the lack of labeled data and high computation demand for segmentation are considered. In this paper, we propose a novel training methodology to train and scale the existing semantic segmentation models for a large number of semantic classes without increasing the memory overhead. In our embedding-based scalable segmentation approach, we reduce the space complexity of the segmentation model's output from O
Ämnesord
- NATURVETENSKAP -- Data- och informationsvetenskap -- Datavetenskap (hsv//swe)
- NATURAL SCIENCES -- Computer and Information Sciences -- Computer Sciences (hsv//eng)
Publikations- och innehållstyp
- ref (ämneskategori)
- kon (ämneskategori)