IRIM at TRECVID 2012: Semantic indexing and instance search

Nicolas Ballas; Benjamin Labbe; Aymen Shabou; Hervé Le Borgne; Philippe Gosselin; Miriam Redi; Bernard Merialdo; Herve Jegou; Jonathan Delhumeau; Rémi Vieux; Boris Mansencal; Jenny Benois-Pineau; Stephane Ayache; Abdelkader Haadi; Bahjat Safadi; Franck Thollard; Nadia Derbas; Georges Quenot; Herve Bredin; Matthieu Cord; Boyang Gao; Chao Zhu; Yuxing Tang; Emmanuel Dellandreav; Charles Edmond Bichot; Liming Chen; Alexandre Benoît; Patrick Lambert; Tiberius Strat; Joseph Razik; Sebastion Paris; Herve Glotin; Tran Ngoc Trung; Dijana Petrovska; Geerard Chollet; Andrei Stoian; Michel Crucianu

IRIM at TRECVID 2012: Semantic indexing and instance search

Nicolas Ballas, Benjamin Labbe, Aymen Shabou, Hervé Le Borgne, Philippe Gosselin, Miriam Redi, Bernard Merialdo, Herve Jegou, Jonathan Delhumeau, Rémi Vieux, Boris Mansencal, Jenny Benois-Pineau, Stephane Ayache, Abdelkader Haadi, Bahjat Safadi, Franck Thollard, Nadia Derbas, Georges Quenot, Herve Bredin, Matthieu CordBoyang Gao, Chao Zhu, Yuxing Tang, Emmanuel Dellandreav, Charles Edmond Bichot, Liming Chen, Alexandre Benoît, Patrick Lambert, Tiberius Strat, Joseph Razik, Sebastion Paris, Herve Glotin, Tran Ngoc Trung, Dijana Petrovska, Geerard Chollet, Andrei Stoian, Michel Crucianu

Research output: Contribution to conference › Paper › peer-review

Abstract

The IRIM group is a consortium of French teams working on Multimedia Indexing and Retrieval. This paper describes its participation to the TRECVID 2012 semantic indexing and instance search tasks. For the semantic indexing task, our approach uses a six-stages processing pipelines for computing scores for the likelihood of a video shot to contain a target concept. These scores are then used for producing a ranked list of images or shots that are the most likely to contain the target concept. The pipeline is composed of the following steps: descriptor extraction, descriptor optimization, classification, fusion of descriptor variants, higher-level fusion, and re-ranking. We evaluated a number of different descriptors and tried different fusion strategies. The best IRIM run has a Mean Inferred Average Precision of 0.2378, which ranked us 4th out of 16 participants.

Original language	English
Publication status	Published - 1 Jan 2012
Externally published	Yes
Event	TREC Video Retrieval Evaluation, TRECVID 2012 - Gaithersburg, MD, United States Duration: 26 Nov 2012 → 28 Nov 2012

Conference

Conference	TREC Video Retrieval Evaluation, TRECVID 2012
Country/Territory	United States
City	Gaithersburg, MD
Period	26/11/12 → 28/11/12

Cite this

Ballas, N., Labbe, B., Shabou, A., Le Borgne, H., Gosselin, P., Redi, M., Merialdo, B., Jegou, H., Delhumeau, J., Vieux, R., Mansencal, B., Benois-Pineau, J., Ayache, S., Haadi, A., Safadi, B., Thollard, F., Derbas, N., Quenot, G., Bredin, H., ... Crucianu, M. (2012). IRIM at TRECVID 2012: Semantic indexing and instance search. Paper presented at TREC Video Retrieval Evaluation, TRECVID 2012, Gaithersburg, MD, United States.

@conference{cedc9f16ee9f4de3b01d65d03bb517dd,

title = "IRIM at TRECVID 2012: Semantic indexing and instance search",

abstract = "The IRIM group is a consortium of French teams working on Multimedia Indexing and Retrieval. This paper describes its participation to the TRECVID 2012 semantic indexing and instance search tasks. For the semantic indexing task, our approach uses a six-stages processing pipelines for computing scores for the likelihood of a video shot to contain a target concept. These scores are then used for producing a ranked list of images or shots that are the most likely to contain the target concept. The pipeline is composed of the following steps: descriptor extraction, descriptor optimization, classification, fusion of descriptor variants, higher-level fusion, and re-ranking. We evaluated a number of different descriptors and tried different fusion strategies. The best IRIM run has a Mean Inferred Average Precision of 0.2378, which ranked us 4th out of 16 participants.",

author = "Nicolas Ballas and Benjamin Labbe and Aymen Shabou and \{Le Borgne\}, Herv{\'e} and Philippe Gosselin and Miriam Redi and Bernard Merialdo and Herve Jegou and Jonathan Delhumeau and R{\'e}mi Vieux and Boris Mansencal and Jenny Benois-Pineau and Stephane Ayache and Abdelkader Haadi and Bahjat Safadi and Franck Thollard and Nadia Derbas and Georges Quenot and Herve Bredin and Matthieu Cord and Boyang Gao and Chao Zhu and Yuxing Tang and Emmanuel Dellandreav and Bichot, \{Charles Edmond\} and Liming Chen and Alexandre Beno{\^i}t and Patrick Lambert and Tiberius Strat and Joseph Razik and Sebastion Paris and Herve Glotin and Trung, \{Tran Ngoc\} and Dijana Petrovska and Geerard Chollet and Andrei Stoian and Michel Crucianu",

year = "2012",

month = jan,

day = "1",

language = "English",

note = "TREC Video Retrieval Evaluation, TRECVID 2012 ; Conference date: 26-11-2012 Through 28-11-2012",

}

Ballas, N, Labbe, B, Shabou, A, Le Borgne, H, Gosselin, P, Redi, M, Merialdo, B, Jegou, H, Delhumeau, J, Vieux, R, Mansencal, B, Benois-Pineau, J, Ayache, S, Haadi, A, Safadi, B, Thollard, F, Derbas, N, Quenot, G, Bredin, H, Cord, M, Gao, B, Zhu, C, Tang, Y, Dellandreav, E, Bichot, CE, Chen, L, Benoît, A, Lambert, P, Strat, T, Razik, J, Paris, S, Glotin, H, Trung, TN, Petrovska, D, Chollet, G, Stoian, A & Crucianu, M 2012, 'IRIM at TRECVID 2012: Semantic indexing and instance search', Paper presented at TREC Video Retrieval Evaluation, TRECVID 2012, Gaithersburg, MD, United States, 26/11/12 - 28/11/12.

TY - CONF

T1 - IRIM at TRECVID 2012

T2 - TREC Video Retrieval Evaluation, TRECVID 2012

AU - Ballas, Nicolas

AU - Labbe, Benjamin

AU - Shabou, Aymen

AU - Le Borgne, Hervé

AU - Gosselin, Philippe

AU - Redi, Miriam

AU - Merialdo, Bernard

AU - Jegou, Herve

AU - Delhumeau, Jonathan

AU - Vieux, Rémi

AU - Mansencal, Boris

AU - Benois-Pineau, Jenny

AU - Ayache, Stephane

AU - Haadi, Abdelkader

AU - Safadi, Bahjat

AU - Thollard, Franck

AU - Derbas, Nadia

AU - Quenot, Georges

AU - Bredin, Herve

AU - Cord, Matthieu

AU - Gao, Boyang

AU - Zhu, Chao

AU - Tang, Yuxing

AU - Dellandreav, Emmanuel

AU - Bichot, Charles Edmond

AU - Chen, Liming

AU - Benoît, Alexandre

AU - Lambert, Patrick

AU - Strat, Tiberius

AU - Razik, Joseph

AU - Paris, Sebastion

AU - Glotin, Herve

AU - Trung, Tran Ngoc

AU - Petrovska, Dijana

AU - Chollet, Geerard

AU - Stoian, Andrei

AU - Crucianu, Michel

PY - 2012/1/1

Y1 - 2012/1/1

N2 - The IRIM group is a consortium of French teams working on Multimedia Indexing and Retrieval. This paper describes its participation to the TRECVID 2012 semantic indexing and instance search tasks. For the semantic indexing task, our approach uses a six-stages processing pipelines for computing scores for the likelihood of a video shot to contain a target concept. These scores are then used for producing a ranked list of images or shots that are the most likely to contain the target concept. The pipeline is composed of the following steps: descriptor extraction, descriptor optimization, classification, fusion of descriptor variants, higher-level fusion, and re-ranking. We evaluated a number of different descriptors and tried different fusion strategies. The best IRIM run has a Mean Inferred Average Precision of 0.2378, which ranked us 4th out of 16 participants.

AB - The IRIM group is a consortium of French teams working on Multimedia Indexing and Retrieval. This paper describes its participation to the TRECVID 2012 semantic indexing and instance search tasks. For the semantic indexing task, our approach uses a six-stages processing pipelines for computing scores for the likelihood of a video shot to contain a target concept. These scores are then used for producing a ranked list of images or shots that are the most likely to contain the target concept. The pipeline is composed of the following steps: descriptor extraction, descriptor optimization, classification, fusion of descriptor variants, higher-level fusion, and re-ranking. We evaluated a number of different descriptors and tried different fusion strategies. The best IRIM run has a Mean Inferred Average Precision of 0.2378, which ranked us 4th out of 16 participants.

M3 - Paper

AN - SCOPUS:84905234242

Y2 - 26 November 2012 through 28 November 2012

ER -

IRIM at TRECVID 2012: Semantic indexing and instance search

Abstract

Conference

Fingerprint

Cite this