Combining Information Sources for Video Retrieval

Thijs Westerveld; Tzvetanka Ianeva; Lioudmila Boldareva; Arjen P. de Vries; Djoerd Hiemstra

Combining Information Sources for Video Retrieval

Thijs Westerveld
, Tzvetanka Ianeva
, Lioudmila Boldareva
, Arjen P. de Vries
, Djoerd Hiemstra

Databases (Former)

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

1 Downloads (Pure)

Abstract

The previous video track results demonstrated that it is far from trivial to take advantage of multiple modalities for the video retrieval search task. For almost any query, results based on ASR transcripts have been better than any other run. This year’s main success in our runs is that a combination of ASR and visual performs better than either alone! In addition we experimented with dynamic shot models, combining topic examples, feature extraction and interactive search.

Original language	English
Title of host publication	TREC Video Retrieval Evaluation Online Proceedings 2003
Place of Publication	Gaithersburg, MD, USA
Publisher	National Institute of Standards and Technology
Number of pages	10
Publication status	Published - Nov 2003
Event	TREC Video Retrieval Evaluation, TRECVID 2003 - Gaithersburg, United States Duration: 17 Nov 2003 → 18 Nov 2003

Workshop

Workshop	TREC Video Retrieval Evaluation, TRECVID 2003
Abbreviated title	TRECVID
Country/Territory	United States
City	Gaithersburg
Period	17/11/03 → 18/11/03

Keywords

DB-MMR: MULTIMEDIA RETRIEVAL

Access to Document

http://www-nlpir.nist.gov/projects/tvpubs/tv.pubs.org.html

Cite this

@inproceedings{817121a27d8a45fabc0dad413c1e3803,

title = "Combining Information Sources for Video Retrieval",

abstract = "The previous video track results demonstrated that it is far from trivial to take advantage of multiple modalities for the video retrieval search task. For almost any query, results based on ASR transcripts have been better than any other run. This year{\textquoteright}s main success in our runs is that a combination of ASR and visual performs better than either alone! In addition we experimented with dynamic shot models, combining topic examples, feature extraction and interactive search.",

keywords = "DB-MMR: MULTIMEDIA RETRIEVAL",

author = "Thijs Westerveld and Tzvetanka Ianeva and Lioudmila Boldareva and \{de Vries\}, \{Arjen P.\} and Djoerd Hiemstra",

note = "http://www-nlpir.nist.gov/projects/trecvid/; TREC Video Retrieval Evaluation, TRECVID 2003, TRECVID ; Conference date: 17-11-2003 Through 18-11-2003",

year = "2003",

month = nov,

language = "English",

booktitle = "TREC Video Retrieval Evaluation Online Proceedings 2003",

publisher = "National Institute of Standards and Technology",

address = "United States",

}

Westerveld, T, Ianeva, T, Boldareva, L, de Vries, AP & Hiemstra, D 2003, Combining Information Sources for Video Retrieval. in TREC Video Retrieval Evaluation Online Proceedings 2003. National Institute of Standards and Technology, Gaithersburg, MD, USA, TREC Video Retrieval Evaluation, TRECVID 2003, Gaithersburg, Maryland, United States, 17/11/03. <http://www-nlpir.nist.gov/projects/tvpubs/tv.pubs.org.html>

Combining Information Sources for Video Retrieval. / Westerveld, Thijs; Ianeva, Tzvetanka; Boldareva, Lioudmila et al.
TREC Video Retrieval Evaluation Online Proceedings 2003. Gaithersburg, MD, USA: National Institute of Standards and Technology, 2003.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Combining Information Sources for Video Retrieval

AU - Westerveld, Thijs

AU - Ianeva, Tzvetanka

AU - Boldareva, Lioudmila

AU - de Vries, Arjen P.

AU - Hiemstra, Djoerd

N1 - http://www-nlpir.nist.gov/projects/trecvid/

PY - 2003/11

Y1 - 2003/11

N2 - The previous video track results demonstrated that it is far from trivial to take advantage of multiple modalities for the video retrieval search task. For almost any query, results based on ASR transcripts have been better than any other run. This year’s main success in our runs is that a combination of ASR and visual performs better than either alone! In addition we experimented with dynamic shot models, combining topic examples, feature extraction and interactive search.

AB - The previous video track results demonstrated that it is far from trivial to take advantage of multiple modalities for the video retrieval search task. For almost any query, results based on ASR transcripts have been better than any other run. This year’s main success in our runs is that a combination of ASR and visual performs better than either alone! In addition we experimented with dynamic shot models, combining topic examples, feature extraction and interactive search.

KW - DB-MMR: MULTIMEDIA RETRIEVAL

M3 - Conference contribution

BT - TREC Video Retrieval Evaluation Online Proceedings 2003

PB - National Institute of Standards and Technology

CY - Gaithersburg, MD, USA

T2 - TREC Video Retrieval Evaluation, TRECVID 2003

Y2 - 17 November 2003 through 18 November 2003

ER -

Combining Information Sources for Video Retrieval

Abstract

Workshop

Keywords

Access to Document

Fingerprint

Cite this