Filters








15 Hits in 3.6 sec

Community-driven hierarchical fusion of numerous classifiers: Application to video semantic indexing

Herve Bredin
2012 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
Application to the combination of 90 classifiers in the framework of TRECVid 2010 Semantic Indexing task shows a 30% increase in performance relative to a baseline flat fusion.  ...  EXPERIMENTS In the framework of TRECVid 2010 Semantic Indexing task [1] , NIST and Quaero provided participants with two corpora annotated with 50 concepts: 120k shots for the development set, N = 150k  ...  Experiments driven on the TRECVid 2010 Semantic Indexing task are summarized in Section 4. Finally, Section 5 concludes the paper.  ... 
doi:10.1109/icassp.2012.6288381 dblp:conf/icassp/Bredin12 fatcat:hkaxesu7jffexfbz5lawokawtm

[Invited Paper] TRECVid Semantic Indexing of Video: A 6-Year Retrospective

George Awad, Cees G. M. Snoek, Alan F. Smeaton, Georges Quénot
2016 ITE Transactions on Media Technology and Applications  
The Semantic Indexing task has been run at TRECVid from 2010 to 2015 with the support of NIST and the Quaero project.  ...  As with the previous High-Level Feature detection task which ran from 2002 to 2009, the semantic indexing task aims at evaluating methods and systems for detecting visual, auditory or multi-modal concepts  ...  The authors wish to thank Paul Over from NIST, now retired, for his work in setting up the TRECVid benchmark and for his help in managing the semantic indexing task.  ... 
doi:10.3169/mta.4.187 fatcat:hcibmza4gzf3naiuvxwop7wgjm

Temporal re-scoring vs. temporal descriptors for semantic indexing of videos

Abdelkader Hamadi, Philippe Mulhem, Georges Quenot
2015 2015 13th International Workshop on Content-Based Multimedia Indexing (CBMI)  
In this paper, we present a comparative study between two methods exploiting the temporal context for semantic video indexing.  ...  Our experiments on TRECVID'12 collection showed interesting results that confirm the usefulness of the temporal context and demonstrate which of the two approaches is more effective.  ...  ACKNOWLEDGEMENTS This work was partly realized as part of the Quaero Program funded by OSEO, French State agency for innovation.  ... 
doi:10.1109/cbmi.2015.7153626 dblp:conf/cbmi/HamadiMQ15 fatcat:y6zfu6pimregjcoii7msa7abnu

Training-free indexing refinement for visual media via multi-semantics

Peng Wang, Lifeng Sun, Shiqiang Yang, Alan F. Smeaton
2017 Neurocomputing  
* factorization and neighbourhood-based graph propagation are both used in the refinement of semantics.  ...  Extended experiments on two heterogeneous datasets, images from wearable cameras and videos from TRECVid, demonstrate the efficacy of the proposed TFR solution.  ...  made directly by the 160 participants of TRECVid or by the annotators of the Quaero project and a 161 total of 28,864,844 annotations was obtained by propagating the initial an-162 notations using the  ... 
doi:10.1016/j.neucom.2016.08.107 fatcat:u33uvlfc2bclrkyckr4c6mqmu4

Two-layers re-ranking approach based on contextual information for visual concepts detection in videos

Abdelkader Hamadi, Georges Quenot, Philippe Mulhem
2012 2012 10th International Workshop on Content-Based Multimedia Indexing (CBMI)  
For concept-based video indexing and retrieval, this idea seems a priori valid.  ...  Experiments conducted on TRECVID 2010 data show that the proposed approach always improves over initial results obtained using either MSVM or KNN classifiers or their late fusion, achieving relative gains  ...  This work was partly realized as part of the Quaero Programme funded by OSEO, French State agency for innovation.  ... 
doi:10.1109/cbmi.2012.6269837 dblp:conf/cbmi/HamadiQM12 fatcat:ib6eidadtzhltl2nv37a6fnsve

Hierarchical Late Fusion for Concept Detection in Videos [chapter]

Sabin Tiberius Strat, Alexandre Benoit, Hervé Bredin, Georges Quénot, Patrick Lambert
2012 Lecture Notes in Computer Science  
Acknowledgments This work was supported by the Quaero Program and the QCompere project, respectively funded by OSEO (French State agency for innovation) and ANR (French national research agency).  ...  Introduction Semantic indexing, as defined in the TRECVid evaluation campaign, consists in automatically detecting the presence of visual concepts in pre-segmented video shots [1] and returning a ranked  ...  Conclusion and future work In this paper, we proposed several ways of combining dozens of input classifiers into better ones, and applied them in the context of the TRECVid 2011 Semantic Indexing task.  ... 
doi:10.1007/978-3-642-33885-4_34 fatcat:7prebtsvvbhtzgrs4bx2deetc4

A PUBLIC AUDIO IDENTIFICATION EVALUATION FRAMEWORK FOR BROADCAST MONITORING

Mathieu Ramona, Sébastien Fenet, Raphaël Blouet, Hervé Bredin, Thomas Fillon, Geoffroy Peeters
2012 Applied Artificial Intelligence  
An example of an application of this framework is finally detailed, that took place during the evaluation campaign of the Quaero project.  ...  Although the domain of audio identification is very active, both in the industry and the academic world, there is at present no common basis to compare the proposed techniques.  ...  Downloaded by [Mathieu Ramona] at 07:54 09 February 2012 M.Ramona et al.  ... 
doi:10.1080/08839514.2012.629840 fatcat:jfeqicdwondx5odxihwnrvhviu

A Local Temporal Context-Based Approach for TV News Story Segmentation

Emilie Dumont, Georges Quenot
2012 2012 IEEE International Conference on Multimedia and Expo  
We investigate the system by experiments conducted using TRECVID 2003 data and protocol of the story boundary detection task and we show that the extension of multimodal descriptors by a local temporal  ...  ACKNOWLEDGMENTS This work was realized as part of the Quaero Programme funded by OSEO, French State agency for innovation.  ...  However, in TRECVID development set, there is 94.1% story boundary near a shot boundary. This means that a system working at the shot level cannot find about 6% of the story boundaries.  ... 
doi:10.1109/icme.2012.3 dblp:conf/icmcs/DumontQ12 fatcat:if5utnqfmnduxasrlh36heotlq

Active Cleaning for Video Corpus Annotation [chapter]

Bahjat Safadi, Stéphane Ayache, Georges Quénot
2012 Lecture Notes in Computer Science  
We have evaluated the actual impact of the active cleaning approach on TRECVID 2007 collection.  ...  In this paper, we have described the active cleaning approach that was used to complement the active learning approach in the TRECVID collaborative annotation.  ...  Automatic concept indexing has been one of the main focus of the TRECVID campaigns (evaluation of video retrieval systems, [12] ) since 2002.  ... 
doi:10.1007/978-3-642-27355-1_48 fatcat:lfpzcg6zhbc7zkoqun4z2wzgvu

Automatic Story Segmentation for TV News Video Using Multiple Modalities

Émilie Dumont, Georges Quénot
2012 International Journal of Digital Multimedia Broadcasting  
We investigate the system by experiments conducted using TRECVID 2003 data and protocol of the story boundary detection task, and we show that the proposed approach outperforms the state-of-the-art methods  ...  Acknowledgment This work was realized as part of the Quaero Programme funded by OSEO, French state agency for innovation.  ...  This system obtained an overall of recall/precision of 0.819/0.851 at the TRECVID 2003 evaluation campaign.  ... 
doi:10.1155/2012/732514 fatcat:5zn6jbhjszdexfxr4gikvlolda

Learned features versus engineered features for multimedia indexing

Mateusz Budnik, Efrain-Leonardo Gutierrez-Gomez, Bahjat Safadi, Denis Pellerin, Georges Quénot
2016 Multimedia tools and applications  
The comparison was made in the context of the TRECVid semantic indexing task.  ...  Learned (or semantic) features are obtained by training classifiers on a source collection containing samples annotated with concepts.  ...  Experiments have been carried out in the context of the semantic indexing task at TRECVid [33, 19] .  ... 
doi:10.1007/s11042-016-4240-2 fatcat:azrhyxhvm5ctxf3xkytwkrtn64

Descriptor optimization for multimedia indexing and retrieval

Bahjat Safadi, Georges Quenot
2013 2013 11th International Workshop on Content-Based Multimedia Indexing (CBMI)  
It has then be applied at large scale for the TRECVid 2012 semantic indexing task on tens of descriptors of various types and with initial dimensionalities from 15 up to 32,768.  ...  The method has been validated and evaluated for a variety of descriptors using TRECVid 2010 semantic indexing task data.  ...  ACKNOWLEDGEMENTS This work was partly realized as part of the Quaero Program funded by OSEO, French State agency for innovation.  ... 
doi:10.1109/cbmi.2013.6576554 dblp:conf/cbmi/SafadiQ13 fatcat:rtxrxdf42ffzbbuj43jjwmchu4

Descriptor optimization for multimedia indexing and retrieval

Bahjat Safadi, Nadia Derbas, Georges Quénot
2014 Multimedia tools and applications  
It has then be applied at large scale for the TRECVid 2012 semantic indexing task on tens of descriptors of various types and with initial dimensionalities from 15 up to 32,768.  ...  The method has been validated and evaluated for a variety of descriptors using TRECVid 2010 semantic indexing task data.  ...  ACKNOWLEDGEMENTS This work was partly realized as part of the Quaero Program funded by OSEO, French State agency for innovation.  ... 
doi:10.1007/s11042-014-2071-6 fatcat:ccjbhngznrefvlqualfrbpy5ci

A factorized model for multiple SVM and multi-label classification for large scale multimedia indexing

Bahjat Safadi, Georges Quenot
2015 2015 13th International Workshop on Content-Based Multimedia Indexing (CBMI)  
Experiments were conducted on a large-scale dataset, namely TRECVid 2012 semantic indexing task.  ...  The proposed method is particularly suited for the detection of many target concepts at once and for highly imbalanced classes (very infrequent concepts).  ...  ACKNOWLEDGEMENTS This work was partly realized as part of the Quaero Program funded by OSEO, French State agency for innovation.  ... 
doi:10.1109/cbmi.2015.7153610 dblp:conf/cbmi/SafadiQ15 fatcat:yoviodf2ircldl73zkdd54gwqa

Topic segmentation of TV-streams by watershed transform and vectorization

Vincent Claveau, Sébastien Lefèvre
2015 Computer Speech and Language  
This work was partially funded by OSEO, French state agency for innovation, in the framework of the Quaero project (www.quaero.org), and by Inria (www.inria.fr).  ...  As it was already identified by several authors (Merlino et al., 1997; Zhai et al., 2005; Poulisse and Moens, 2009; Guinaudeau et al., 2012; Dumont and Quénot, 2012) , additional clues are available in  ...  For instance, segmentation of TV news reports have been explored in TRECVID competition 2 (Amir et al., 2004 , for a representative multimodal system).  ... 
doi:10.1016/j.csl.2014.04.006 fatcat:s6ewbijwbjdzte4lsi7ia6bh74