Small Vocabulary with Saliency Matching for Video Copy Detection

Huamin Ren, Thomas B. Moeslund, Sheng Tang, Heri Ramampiaro

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

332 Downloads (Pure)

Abstract

The importance of copy detection has led to a substantial amount of research in recent years, among which Bag of visual Words (BoW) plays an important role due to its ability to effectively handling occlusion and some minor transformations. One crucial issue in BoW approaches is the size of vocabulary. BoW descriptors under a small vocabulary can be both robust and efficient, while keeping high recall rate compared with large vocabulary. However, the high false positives exists in small vocabulary also limits its application. To address this problem in small vocabulary, we propose a novel matching algorithm based on salient visual words selection. More specifically, the variation of visual words across a given video are represented as trajectories and those containing locally asymptotically stable points are selected as salient visual words. Then we attempt to measure the similarity of two videos through saliency matching merely based on the selected salient visual words to remove false positives. Our experiments show that a small codebook with saliency matching is quite competitive in video copy detection. With the incorporation of the proposed saliency matching, the precision can be improved by 30% on average compared with the state-of-the-art technique. Moreover, our proposed method is capable of detecting severe transformations, e.g. picture in picture and post production.
Original languageEnglish
Title of host publicationVISAPP 2013 - Proceedings of the International Conference on Computer Vision Theory and Applications : The International Conference on Computer Vision Theory and Applications
Number of pages6
PublisherInstitute for Systems and Technologies of Information, Control and Communication
Publication date2013
Pages768-773
ISBN (Print)978-989856547-1
Publication statusPublished - 2013
EventVISAPP 2013 - Barcelona, Spain
Duration: 21 Feb 201324 Feb 2013
Conference number: 8
http://www.visapp.visigrapp.org/?y=2013

Conference

ConferenceVISAPP 2013
Number8
CountrySpain
CityBarcelona
Period21/02/201324/02/2013
Internet address

Fingerprint

Trajectories
Experiments

Keywords

  • Visual Codebook; Bag of Words; Saliency Matching; Copy Detection.

Cite this

Ren, H., Moeslund, T. B., Tang, S., & Ramampiaro, H. (2013). Small Vocabulary with Saliency Matching for Video Copy Detection. In VISAPP 2013 - Proceedings of the International Conference on Computer Vision Theory and Applications: The International Conference on Computer Vision Theory and Applications (pp. 768-773). Institute for Systems and Technologies of Information, Control and Communication.
Ren, Huamin ; Moeslund, Thomas B. ; Tang, Sheng ; Ramampiaro, Heri. / Small Vocabulary with Saliency Matching for Video Copy Detection. VISAPP 2013 - Proceedings of the International Conference on Computer Vision Theory and Applications: The International Conference on Computer Vision Theory and Applications. Institute for Systems and Technologies of Information, Control and Communication, 2013. pp. 768-773
@inproceedings{8cedf71fec3a4dedaa74528084c100c3,
title = "Small Vocabulary with Saliency Matching for Video Copy Detection",
abstract = "The importance of copy detection has led to a substantial amount of research in recent years, among which Bag of visual Words (BoW) plays an important role due to its ability to effectively handling occlusion and some minor transformations. One crucial issue in BoW approaches is the size of vocabulary. BoW descriptors under a small vocabulary can be both robust and efficient, while keeping high recall rate compared with large vocabulary. However, the high false positives exists in small vocabulary also limits its application. To address this problem in small vocabulary, we propose a novel matching algorithm based on salient visual words selection. More specifically, the variation of visual words across a given video are represented as trajectories and those containing locally asymptotically stable points are selected as salient visual words. Then we attempt to measure the similarity of two videos through saliency matching merely based on the selected salient visual words to remove false positives. Our experiments show that a small codebook with saliency matching is quite competitive in video copy detection. With the incorporation of the proposed saliency matching, the precision can be improved by 30{\%} on average compared with the state-of-the-art technique. Moreover, our proposed method is capable of detecting severe transformations, e.g. picture in picture and post production.",
keywords = "Visual Codebook; Bag of Words; Saliency Matching; Copy Detection.",
author = "Huamin Ren and Moeslund, {Thomas B.} and Sheng Tang and Heri Ramampiaro",
year = "2013",
language = "English",
isbn = "978-989856547-1",
pages = "768--773",
booktitle = "VISAPP 2013 - Proceedings of the International Conference on Computer Vision Theory and Applications",
publisher = "Institute for Systems and Technologies of Information, Control and Communication",

}

Ren, H, Moeslund, TB, Tang, S & Ramampiaro, H 2013, Small Vocabulary with Saliency Matching for Video Copy Detection. in VISAPP 2013 - Proceedings of the International Conference on Computer Vision Theory and Applications: The International Conference on Computer Vision Theory and Applications. Institute for Systems and Technologies of Information, Control and Communication, pp. 768-773, VISAPP 2013, Barcelona, Spain, 21/02/2013.

Small Vocabulary with Saliency Matching for Video Copy Detection. / Ren, Huamin; Moeslund, Thomas B.; Tang, Sheng; Ramampiaro, Heri.

VISAPP 2013 - Proceedings of the International Conference on Computer Vision Theory and Applications: The International Conference on Computer Vision Theory and Applications. Institute for Systems and Technologies of Information, Control and Communication, 2013. p. 768-773.

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Small Vocabulary with Saliency Matching for Video Copy Detection

AU - Ren, Huamin

AU - Moeslund, Thomas B.

AU - Tang, Sheng

AU - Ramampiaro, Heri

PY - 2013

Y1 - 2013

N2 - The importance of copy detection has led to a substantial amount of research in recent years, among which Bag of visual Words (BoW) plays an important role due to its ability to effectively handling occlusion and some minor transformations. One crucial issue in BoW approaches is the size of vocabulary. BoW descriptors under a small vocabulary can be both robust and efficient, while keeping high recall rate compared with large vocabulary. However, the high false positives exists in small vocabulary also limits its application. To address this problem in small vocabulary, we propose a novel matching algorithm based on salient visual words selection. More specifically, the variation of visual words across a given video are represented as trajectories and those containing locally asymptotically stable points are selected as salient visual words. Then we attempt to measure the similarity of two videos through saliency matching merely based on the selected salient visual words to remove false positives. Our experiments show that a small codebook with saliency matching is quite competitive in video copy detection. With the incorporation of the proposed saliency matching, the precision can be improved by 30% on average compared with the state-of-the-art technique. Moreover, our proposed method is capable of detecting severe transformations, e.g. picture in picture and post production.

AB - The importance of copy detection has led to a substantial amount of research in recent years, among which Bag of visual Words (BoW) plays an important role due to its ability to effectively handling occlusion and some minor transformations. One crucial issue in BoW approaches is the size of vocabulary. BoW descriptors under a small vocabulary can be both robust and efficient, while keeping high recall rate compared with large vocabulary. However, the high false positives exists in small vocabulary also limits its application. To address this problem in small vocabulary, we propose a novel matching algorithm based on salient visual words selection. More specifically, the variation of visual words across a given video are represented as trajectories and those containing locally asymptotically stable points are selected as salient visual words. Then we attempt to measure the similarity of two videos through saliency matching merely based on the selected salient visual words to remove false positives. Our experiments show that a small codebook with saliency matching is quite competitive in video copy detection. With the incorporation of the proposed saliency matching, the precision can be improved by 30% on average compared with the state-of-the-art technique. Moreover, our proposed method is capable of detecting severe transformations, e.g. picture in picture and post production.

KW - Visual Codebook; Bag of Words; Saliency Matching; Copy Detection.

UR - http://www.scopus.com/inward/record.url?scp=84878257645&partnerID=8YFLogxK

M3 - Article in proceeding

SN - 978-989856547-1

SP - 768

EP - 773

BT - VISAPP 2013 - Proceedings of the International Conference on Computer Vision Theory and Applications

PB - Institute for Systems and Technologies of Information, Control and Communication

ER -

Ren H, Moeslund TB, Tang S, Ramampiaro H. Small Vocabulary with Saliency Matching for Video Copy Detection. In VISAPP 2013 - Proceedings of the International Conference on Computer Vision Theory and Applications: The International Conference on Computer Vision Theory and Applications. Institute for Systems and Technologies of Information, Control and Communication. 2013. p. 768-773