Detection of multiple instances of video objects

Andrei Bursuc; Titus Zaharia; Françoise Prêteux

doi:10.1109/SITIS.2011.74

Detection of multiple instances of video objects

Andrei Bursuc, Titus Zaharia, Françoise Prêteux

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

Abstract

This paper tackles the issue of retrieving different instances of an object of interest within a given video document or in a video database. The principle consists of considering a semi-global image representation based on an over-segmentation of image frames. An aggregation mechanism is then applied in order to group a set of sub-regions into an object similar to the query, under a global similarity criterion. Two different types of approaches are proposed. The first one involves a greedy, dynamic region construction method. The second is based on simulated annealing, and aims at determining a global optimum of the similarity function. Experimental results show promising performances, with FT and BE detection rates of up to 66% and 86%, respectively.

Original language	English
Title of host publication	Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011
Pages	446-453
Number of pages	8
DOIs	https://doi.org/10.1109/SITIS.2011.74
Publication status	Published - 1 Dec 2011
Externally published	Yes
Event	7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011 - Dijon, France Duration: 28 Nov 2011 → 1 Dec 2011

Publication series

Name	Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011

Conference

Conference	7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011
Country/Territory	France
City	Dijon
Period	28/11/11 → 1/12/11

Keywords

MPEG-7 visual descriptors
Multiple instance detection
Object-based indexing and retrieval
Partial matching
Video indexing

Access to Document

10.1109/SITIS.2011.74

Cite this

Bursuc, A., Zaharia, T., & Prêteux, F. (2011). Detection of multiple instances of video objects. In Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011 (pp. 446-453). Article 6120686 (Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011). https://doi.org/10.1109/SITIS.2011.74

@inproceedings{0e7ddd29f0244c3a88b666133b4a768b,

title = "Detection of multiple instances of video objects",

abstract = "This paper tackles the issue of retrieving different instances of an object of interest within a given video document or in a video database. The principle consists of considering a semi-global image representation based on an over-segmentation of image frames. An aggregation mechanism is then applied in order to group a set of sub-regions into an object similar to the query, under a global similarity criterion. Two different types of approaches are proposed. The first one involves a greedy, dynamic region construction method. The second is based on simulated annealing, and aims at determining a global optimum of the similarity function. Experimental results show promising performances, with FT and BE detection rates of up to 66\% and 86\%, respectively.",

keywords = "MPEG-7 visual descriptors, Multiple instance detection, Object-based indexing and retrieval, Partial matching, Video indexing",

author = "Andrei Bursuc and Titus Zaharia and Fran{\c c}oise Pr{\^e}teux",

year = "2011",

month = dec,

day = "1",

doi = "10.1109/SITIS.2011.74",

language = "English",

isbn = "9780769546353",

series = "Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011",

pages = "446--453",

booktitle = "Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011",

note = "7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011 ; Conference date: 28-11-2011 Through 01-12-2011",

}

Bursuc, A, Zaharia, T & Prêteux, F 2011, Detection of multiple instances of video objects. in Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011., 6120686, Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011, pp. 446-453, 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011, Dijon, France, 28/11/11. https://doi.org/10.1109/SITIS.2011.74

Detection of multiple instances of video objects. / Bursuc, Andrei; Zaharia, Titus; Prêteux, Françoise.
Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011. 2011. p. 446-453 6120686 (Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011).

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

TY - GEN

T1 - Detection of multiple instances of video objects

AU - Bursuc, Andrei

AU - Zaharia, Titus

AU - Prêteux, Françoise

PY - 2011/12/1

Y1 - 2011/12/1

N2 - This paper tackles the issue of retrieving different instances of an object of interest within a given video document or in a video database. The principle consists of considering a semi-global image representation based on an over-segmentation of image frames. An aggregation mechanism is then applied in order to group a set of sub-regions into an object similar to the query, under a global similarity criterion. Two different types of approaches are proposed. The first one involves a greedy, dynamic region construction method. The second is based on simulated annealing, and aims at determining a global optimum of the similarity function. Experimental results show promising performances, with FT and BE detection rates of up to 66% and 86%, respectively.

AB - This paper tackles the issue of retrieving different instances of an object of interest within a given video document or in a video database. The principle consists of considering a semi-global image representation based on an over-segmentation of image frames. An aggregation mechanism is then applied in order to group a set of sub-regions into an object similar to the query, under a global similarity criterion. Two different types of approaches are proposed. The first one involves a greedy, dynamic region construction method. The second is based on simulated annealing, and aims at determining a global optimum of the similarity function. Experimental results show promising performances, with FT and BE detection rates of up to 66% and 86%, respectively.

KW - MPEG-7 visual descriptors

KW - Multiple instance detection

KW - Object-based indexing and retrieval

KW - Partial matching

KW - Video indexing

U2 - 10.1109/SITIS.2011.74

DO - 10.1109/SITIS.2011.74

M3 - Conference contribution

AN - SCOPUS:84855932284

SN - 9780769546353

T3 - Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011

SP - 446

EP - 453

BT - Proceedings - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011

T2 - 7th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2011

Y2 - 28 November 2011 through 1 December 2011

ER -

Detection of multiple instances of video objects

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this