Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | 2015 IEEE International Conference on Image Processing, ICIP 2015 - Proceedings |
| Herausgeber (Verlag) | IEEE Computer Society |
| Seiten | 902-906 |
| Seitenumfang | 5 |
| ISBN (elektronisch) | 9781479983391 |
| Publikationsstatus | Veröffentlicht - 9 Dez. 2015 |
| Veranstaltung | IEEE International Conference on Image Processing, ICIP 2015 - Quebec City, Kanada Dauer: 27 Sept. 2015 → 30 Sept. 2015 |
Publikationsreihe
| Name | Proceedings - International Conference on Image Processing, ICIP |
|---|---|
| Band | 2015-December |
| ISSN (Print) | 1522-4880 |
Abstract
Many recent superpixel algorithms for video content rely on dense optical flow vectors to propagate segmentation results from one frame to the next. In this paper, we assess the impact of the optical flow quality on the over-segmentation quality. Our evaluation shows that it is indispensable for videos with large object displacement and camera motion. But due to the high computational costs high-quality, dense optical flow is not suitable for real-time applications. Therefore, we propose a fast propagation scheme that is based on sparse feature tracking and mesh-based image warping. In a thorough evaluation, we compare our proposed scheme to the results of other state-of-the-art propagation methods using established benchmarks. The results show that our method speeds up the propagation process by a factor of 100 while producing a comparable segmentation quality.
ASJC Scopus Sachgebiete
- Informatik (insg.)
- Software
- Informatik (insg.)
- Maschinelles Sehen und Mustererkennung
- Informatik (insg.)
- Signalverarbeitung
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
2015 IEEE International Conference on Image Processing, ICIP 2015 - Proceedings. IEEE Computer Society, 2015. S. 902-906 7350930 (Proceedings - International Conference on Image Processing, ICIP; Band 2015-December).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - Fast label propagation for real-time superpixels for video content
AU - Reso, Matthias
AU - Jachalsky, Jorn
AU - Rosenhahn, Bodo
AU - Ostermann, Jörn
PY - 2015/12/9
Y1 - 2015/12/9
N2 - Many recent superpixel algorithms for video content rely on dense optical flow vectors to propagate segmentation results from one frame to the next. In this paper, we assess the impact of the optical flow quality on the over-segmentation quality. Our evaluation shows that it is indispensable for videos with large object displacement and camera motion. But due to the high computational costs high-quality, dense optical flow is not suitable for real-time applications. Therefore, we propose a fast propagation scheme that is based on sparse feature tracking and mesh-based image warping. In a thorough evaluation, we compare our proposed scheme to the results of other state-of-the-art propagation methods using established benchmarks. The results show that our method speeds up the propagation process by a factor of 100 while producing a comparable segmentation quality.
AB - Many recent superpixel algorithms for video content rely on dense optical flow vectors to propagate segmentation results from one frame to the next. In this paper, we assess the impact of the optical flow quality on the over-segmentation quality. Our evaluation shows that it is indispensable for videos with large object displacement and camera motion. But due to the high computational costs high-quality, dense optical flow is not suitable for real-time applications. Therefore, we propose a fast propagation scheme that is based on sparse feature tracking and mesh-based image warping. In a thorough evaluation, we compare our proposed scheme to the results of other state-of-the-art propagation methods using established benchmarks. The results show that our method speeds up the propagation process by a factor of 100 while producing a comparable segmentation quality.
KW - optical flow
KW - Superpixel
KW - supervoxel
UR - http://www.scopus.com/inward/record.url?scp=84956658538&partnerID=8YFLogxK
U2 - 10.1109/icip.2015.7350930
DO - 10.1109/icip.2015.7350930
M3 - Conference contribution
AN - SCOPUS:84956658538
T3 - Proceedings - International Conference on Image Processing, ICIP
SP - 902
EP - 906
BT - 2015 IEEE International Conference on Image Processing, ICIP 2015 - Proceedings
PB - IEEE Computer Society
T2 - IEEE International Conference on Image Processing, ICIP 2015
Y2 - 27 September 2015 through 30 September 2015
ER -