Retrieval of multiple instances of objects in videos

Andrei Bursuc; Titus Zaharia; Françoise Prêteux

doi:10.1007/978-3-642-27355-1_34

Retrieval of multiple instances of objects in videos

Andrei Bursuc
, Titus Zaharia
, Françoise Prêteux

CNRS SAMOVAR UMR 5157
Bell Labs
Mines ParisTech

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

3 Citations (Scopus)

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 in 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 strategies 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. Experimental results show promising performances, with object detection rates of up to 79%.

Original language	English
Title of host publication	Advances in Multimedia Modeling - 18th International Conference, MMM 2012, Proceedings
Pages	358-369
Number of pages	12
DOIs	https://doi.org/10.1007/978-3-642-27355-1_34
Publication status	Published - 18 Jan 2012
Externally published	Yes
Event	18th International Conference on Multimedia Modeling, MMM 2012 - Klagenfurt, Austria Duration: 4 Jan 2012 → 6 Jan 2012

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	7131 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	18th International Conference on Multimedia Modeling, MMM 2012
Country/Territory	Austria
City	Klagenfurt
Period	4/01/12 → 6/01/12

Keywords

MPEG-7 visual descriptors
multiple instance detection
object-based indexing and retrieval
partial matching
video indexing

Access to Document

10.1007/978-3-642-27355-1_34

Cite this

Bursuc, A., Zaharia, T., & Prêteux, F. (2012). Retrieval of multiple instances of objects in videos. In Advances in Multimedia Modeling - 18th International Conference, MMM 2012, Proceedings (pp. 358-369). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7131 LNCS). https://doi.org/10.1007/978-3-642-27355-1_34

@inproceedings{3d13c24139bb43bb9954b386a6b9e099,

title = "Retrieval of multiple instances of objects in videos",

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 in 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 strategies 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. Experimental results show promising performances, with object detection rates of up to 79\%.",

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 = "2012",

month = jan,

day = "18",

doi = "10.1007/978-3-642-27355-1\_34",

language = "English",

isbn = "9783642273544",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "358--369",

booktitle = "Advances in Multimedia Modeling - 18th International Conference, MMM 2012, Proceedings",

note = "18th International Conference on Multimedia Modeling, MMM 2012 ; Conference date: 04-01-2012 Through 06-01-2012",

}

Bursuc, A, Zaharia, T & Prêteux, F 2012, Retrieval of multiple instances of objects in videos. in Advances in Multimedia Modeling - 18th International Conference, MMM 2012, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7131 LNCS, pp. 358-369, 18th International Conference on Multimedia Modeling, MMM 2012, Klagenfurt, Austria, 4/01/12. https://doi.org/10.1007/978-3-642-27355-1_34

Retrieval of multiple instances of objects in videos. / Bursuc, Andrei; Zaharia, Titus; Prêteux, Françoise.
Advances in Multimedia Modeling - 18th International Conference, MMM 2012, Proceedings. 2012. p. 358-369 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7131 LNCS).

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

TY - GEN

T1 - Retrieval of multiple instances of objects in videos

AU - Bursuc, Andrei

AU - Zaharia, Titus

AU - Prêteux, Françoise

PY - 2012/1/18

Y1 - 2012/1/18

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 in 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 strategies 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. Experimental results show promising performances, with object detection rates of up to 79%.

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 in 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 strategies 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. Experimental results show promising performances, with object detection rates of up to 79%.

KW - MPEG-7 visual descriptors

KW - multiple instance detection

KW - object-based indexing and retrieval

KW - partial matching

KW - video indexing

U2 - 10.1007/978-3-642-27355-1_34

DO - 10.1007/978-3-642-27355-1_34

M3 - Conference contribution

AN - SCOPUS:84855772526

SN - 9783642273544

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 358

EP - 369

BT - Advances in Multimedia Modeling - 18th International Conference, MMM 2012, Proceedings

T2 - 18th International Conference on Multimedia Modeling, MMM 2012

Y2 - 4 January 2012 through 6 January 2012

ER -

Retrieval of multiple instances of objects in videos

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this