University of Marburg at TRECVID 2007: Shot boundary detection and high level feature extraction

Research output: Contribution to conferencePaperResearchpeer review

Authors

  • Markus Mühling
  • Ralph Ewerth
  • Thilo Stadelmann
  • Christian Zöfel
  • Bing Shi
  • Bernd Freisleben

External Research Organisations

  • University of Siegen
  • Philipps-Universität Marburg
View graph of relations

Details

Original languageEnglish
Publication statusPublished - 2007
Externally publishedYes
EventTREC Video Retrieval Evaluation, TRECVID 2007 - Gaithersburg, MD, United States
Duration: 5 Nov 20076 Nov 2007

Conference

ConferenceTREC Video Retrieval Evaluation, TRECVID 2007
Country/TerritoryUnited States
CityGaithersburg, MD
Period5 Nov 20076 Nov 2007

Abstract

In this paper, we summarize our results for the shot boundary and high level feature detection task at TRECVID 2007. Our shot boundary detection approach of previous TRECVID evaluations served as a basis for our experiments this year and was modified in several ways. First, we have incorporated a new metric selection for cut detection based on the evaluation of a clustering result. Second, we have tested the possibility to improve cut detection results via self-supervised learning. Third, the unsupervised approach for gradual transition detection has been supplemented with a false alarm removal method using a state-of-the art camera motion estimation approach. Regarding high-level feature detection, one focus of this year's task was to investigate the question how well a trained system generalizes from the TRECVID 2005 news data to this year's Sound and Vision data. However, only two institutes have submitted four runs of the related type "a" for evaluation (three of them were submitted by us). In this paper, we present our experiments for the high-level feature task with respect to the generalization capabilities of our system trained on broadcast news videos. For this purpose, we have conducted several experiments using our system which is based on low-level features as well as on state-ofthe- art approaches for camera motion estimation, text detection, face detection and audio segmentation.

ASJC Scopus subject areas

Cite this

University of Marburg at TRECVID 2007: Shot boundary detection and high level feature extraction. / Mühling, Markus; Ewerth, Ralph; Stadelmann, Thilo et al.
2007. Paper presented at TREC Video Retrieval Evaluation, TRECVID 2007, Gaithersburg, MD, United States.

Research output: Contribution to conferencePaperResearchpeer review

Mühling, M, Ewerth, R, Stadelmann, T, Zöfel, C, Shi, B & Freisleben, B 2007, 'University of Marburg at TRECVID 2007: Shot boundary detection and high level feature extraction', Paper presented at TREC Video Retrieval Evaluation, TRECVID 2007, Gaithersburg, MD, United States, 5 Nov 2007 - 6 Nov 2007.
Mühling, M., Ewerth, R., Stadelmann, T., Zöfel, C., Shi, B., & Freisleben, B. (2007). University of Marburg at TRECVID 2007: Shot boundary detection and high level feature extraction. Paper presented at TREC Video Retrieval Evaluation, TRECVID 2007, Gaithersburg, MD, United States.
Mühling M, Ewerth R, Stadelmann T, Zöfel C, Shi B, Freisleben B. University of Marburg at TRECVID 2007: Shot boundary detection and high level feature extraction. 2007. Paper presented at TREC Video Retrieval Evaluation, TRECVID 2007, Gaithersburg, MD, United States.
Mühling, Markus ; Ewerth, Ralph ; Stadelmann, Thilo et al. / University of Marburg at TRECVID 2007 : Shot boundary detection and high level feature extraction. Paper presented at TREC Video Retrieval Evaluation, TRECVID 2007, Gaithersburg, MD, United States.
Download
@conference{d912654ce1fb4375ac7760f1c826770d,
title = "University of Marburg at TRECVID 2007: Shot boundary detection and high level feature extraction",
abstract = "In this paper, we summarize our results for the shot boundary and high level feature detection task at TRECVID 2007. Our shot boundary detection approach of previous TRECVID evaluations served as a basis for our experiments this year and was modified in several ways. First, we have incorporated a new metric selection for cut detection based on the evaluation of a clustering result. Second, we have tested the possibility to improve cut detection results via self-supervised learning. Third, the unsupervised approach for gradual transition detection has been supplemented with a false alarm removal method using a state-of-the art camera motion estimation approach. Regarding high-level feature detection, one focus of this year's task was to investigate the question how well a trained system generalizes from the TRECVID 2005 news data to this year's Sound and Vision data. However, only two institutes have submitted four runs of the related type {"}a{"} for evaluation (three of them were submitted by us). In this paper, we present our experiments for the high-level feature task with respect to the generalization capabilities of our system trained on broadcast news videos. For this purpose, we have conducted several experiments using our system which is based on low-level features as well as on state-ofthe- art approaches for camera motion estimation, text detection, face detection and audio segmentation.",
author = "Markus M{\"u}hling and Ralph Ewerth and Thilo Stadelmann and Christian Z{\"o}fel and Bing Shi and Bernd Freisleben",
year = "2007",
language = "English",
note = "TREC Video Retrieval Evaluation, TRECVID 2007 ; Conference date: 05-11-2007 Through 06-11-2007",

}

Download

TY - CONF

T1 - University of Marburg at TRECVID 2007

T2 - TREC Video Retrieval Evaluation, TRECVID 2007

AU - Mühling, Markus

AU - Ewerth, Ralph

AU - Stadelmann, Thilo

AU - Zöfel, Christian

AU - Shi, Bing

AU - Freisleben, Bernd

PY - 2007

Y1 - 2007

N2 - In this paper, we summarize our results for the shot boundary and high level feature detection task at TRECVID 2007. Our shot boundary detection approach of previous TRECVID evaluations served as a basis for our experiments this year and was modified in several ways. First, we have incorporated a new metric selection for cut detection based on the evaluation of a clustering result. Second, we have tested the possibility to improve cut detection results via self-supervised learning. Third, the unsupervised approach for gradual transition detection has been supplemented with a false alarm removal method using a state-of-the art camera motion estimation approach. Regarding high-level feature detection, one focus of this year's task was to investigate the question how well a trained system generalizes from the TRECVID 2005 news data to this year's Sound and Vision data. However, only two institutes have submitted four runs of the related type "a" for evaluation (three of them were submitted by us). In this paper, we present our experiments for the high-level feature task with respect to the generalization capabilities of our system trained on broadcast news videos. For this purpose, we have conducted several experiments using our system which is based on low-level features as well as on state-ofthe- art approaches for camera motion estimation, text detection, face detection and audio segmentation.

AB - In this paper, we summarize our results for the shot boundary and high level feature detection task at TRECVID 2007. Our shot boundary detection approach of previous TRECVID evaluations served as a basis for our experiments this year and was modified in several ways. First, we have incorporated a new metric selection for cut detection based on the evaluation of a clustering result. Second, we have tested the possibility to improve cut detection results via self-supervised learning. Third, the unsupervised approach for gradual transition detection has been supplemented with a false alarm removal method using a state-of-the art camera motion estimation approach. Regarding high-level feature detection, one focus of this year's task was to investigate the question how well a trained system generalizes from the TRECVID 2005 news data to this year's Sound and Vision data. However, only two institutes have submitted four runs of the related type "a" for evaluation (three of them were submitted by us). In this paper, we present our experiments for the high-level feature task with respect to the generalization capabilities of our system trained on broadcast news videos. For this purpose, we have conducted several experiments using our system which is based on low-level features as well as on state-ofthe- art approaches for camera motion estimation, text detection, face detection and audio segmentation.

UR - http://www.scopus.com/inward/record.url?scp=84905159151&partnerID=8YFLogxK

M3 - Paper

Y2 - 5 November 2007 through 6 November 2007

ER -