Details
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 6th ACM International Conference on Image and Video Retrieval, CIVR 2007 |
| Pages | 210-217 |
| Number of pages | 8 |
| Publication status | Published - 9 Jul 2007 |
| Externally published | Yes |
| Event | 6th ACM International Conference on Image and Video Retrieval, CIVR 2007 - Amsterdam, Netherlands Duration: 9 Jul 2007 → 11 Jul 2007 |
Publication series
| Name | Proceedings of the 6th ACM International Conference on Image and Video Retrieval, CIVR 2007 |
|---|
Abstract
In recent years, many features have been suggested to facilitate the task of video retrieval. However, up to now, three-dimensional (3D) scene information has not been utilized to retrieve shots from a database consisting of monocular video sequences. In this paper, we propose the use of depth features for video retrieval purposes. Our depth feature extraction approach is based on a method that originally has been suggested to scan 3D objects with a single camera exploiting the motion parallax. To increase the number of video shots for which the depth feature extraction method is applicable, we present an extension of the self-calibration algorithm of this method. Furthermore, a depth map representation is presented and an adequate distance measure is suggested to compare depth maps. Finally, the extracted depth features are used to retrieve video shots according to the three-dimensional scene content of a shot. Experimental results for the comprehensive TRECVID 2005 video data set demonstrate the usefulness of the proposed depth features for video retrieval.
Keywords
- 3D features, Depth estimation, Depth features, Monocular videos, Semantic video retrieval, Video indexing, Video retrieval
ASJC Scopus subject areas
- Engineering(all)
- Electrical and Electronic Engineering
- Computer Science(all)
- General Computer Science
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Proceedings of the 6th ACM International Conference on Image and Video Retrieval, CIVR 2007. 2007. p. 210-217 (Proceedings of the 6th ACM International Conference on Image and Video Retrieval, CIVR 2007).
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Using depth features to retrieve monocular video shots
AU - Ewerth, Ralph
AU - Schwalb, Martin
AU - Freisleben, Bernd
PY - 2007/7/9
Y1 - 2007/7/9
N2 - In recent years, many features have been suggested to facilitate the task of video retrieval. However, up to now, three-dimensional (3D) scene information has not been utilized to retrieve shots from a database consisting of monocular video sequences. In this paper, we propose the use of depth features for video retrieval purposes. Our depth feature extraction approach is based on a method that originally has been suggested to scan 3D objects with a single camera exploiting the motion parallax. To increase the number of video shots for which the depth feature extraction method is applicable, we present an extension of the self-calibration algorithm of this method. Furthermore, a depth map representation is presented and an adequate distance measure is suggested to compare depth maps. Finally, the extracted depth features are used to retrieve video shots according to the three-dimensional scene content of a shot. Experimental results for the comprehensive TRECVID 2005 video data set demonstrate the usefulness of the proposed depth features for video retrieval.
AB - In recent years, many features have been suggested to facilitate the task of video retrieval. However, up to now, three-dimensional (3D) scene information has not been utilized to retrieve shots from a database consisting of monocular video sequences. In this paper, we propose the use of depth features for video retrieval purposes. Our depth feature extraction approach is based on a method that originally has been suggested to scan 3D objects with a single camera exploiting the motion parallax. To increase the number of video shots for which the depth feature extraction method is applicable, we present an extension of the self-calibration algorithm of this method. Furthermore, a depth map representation is presented and an adequate distance measure is suggested to compare depth maps. Finally, the extracted depth features are used to retrieve video shots according to the three-dimensional scene content of a shot. Experimental results for the comprehensive TRECVID 2005 video data set demonstrate the usefulness of the proposed depth features for video retrieval.
KW - 3D features
KW - Depth estimation
KW - Depth features
KW - Monocular videos
KW - Semantic video retrieval
KW - Video indexing
KW - Video retrieval
UR - http://www.scopus.com/inward/record.url?scp=36849040065&partnerID=8YFLogxK
U2 - 10.1145/1282280.1282315
DO - 10.1145/1282280.1282315
M3 - Conference contribution
AN - SCOPUS:36849040065
SN - 1595937331
SN - 9781595937339
T3 - Proceedings of the 6th ACM International Conference on Image and Video Retrieval, CIVR 2007
SP - 210
EP - 217
BT - Proceedings of the 6th ACM International Conference on Image and Video Retrieval, CIVR 2007
T2 - 6th ACM International Conference on Image and Video Retrieval, CIVR 2007
Y2 - 9 July 2007 through 11 July 2007
ER -