SlideShare une entreprise Scribd logo
1  sur  21
Les enjeux scientifiques de l’indexation vidéo Patrick Gros  Responsable de l’équipe TEXMEX INRIA Rennes et IRISA http://www.irisa.fr/texmex
Qu’est ce que l’indexation vidéo ? ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]
Des applications ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]
Quelques opérations ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]
Des opérations de base ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]
Mais… ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]
Les problèmes scientifiques ,[object Object],[object Object],[object Object],[object Object]
La temporalité ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]
Exemple des Modèles de Markov cachés ,[object Object],[object Object],[object Object],[object Object]
Exemple des Modèles de Markov cachés ,[object Object],[object Object],[object Object],[object Object]
Exemple des Modèles de Markov cachés ,[object Object],[object Object],[object Object],[object Object],[object Object]
La généricité ,[object Object],[object Object],[object Object]
La généricité ,[object Object],[object Object],[object Object],[object Object]
La généricité ,[object Object],[object Object],[object Object]
La multimodalité ,[object Object],[object Object],[object Object]
La sémantique ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]
La sémantique ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]
Adaptation dynamique ,[object Object],[object Object],€ ASR system ... receives a single electoral vote in this state Un flux long (audio)‏ Hypothèse de transcription (texte)‏ ... ...
Web-based topic adaptation ... ... € … thus a  candidate  who fails to carry a particular  state  receives not a single  electoral   vote  in that  state  for the popular votes received since residential  elections  are won by  electoral  ... candidate state election 3.  Building of an adaptation corpus candidate vote electoral vote 2.  Querying 1.  Keyword spotting Adaptation LM 4.a  Training of a topic-specific LM   4.b  Mix of this LM and the general one Baseline LM + Adapted LM = Web search engine ✘ ✔ ✔ ✔ ✘ ✔ ✘ ✔
La sémantique ,[object Object],[object Object]
Conclusion ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]

Contenu connexe

En vedette

Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
gt_ebuddy
 
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
gt_ebuddy
 
Feature Extraction
Feature ExtractionFeature Extraction
Feature Extraction
skylian
 
Expectation Maximization and Gaussian Mixture Models
Expectation Maximization and Gaussian Mixture ModelsExpectation Maximization and Gaussian Mixture Models
Expectation Maximization and Gaussian Mixture Models
petitegeek
 
K-means, EM and Mixture models
K-means, EM and Mixture modelsK-means, EM and Mixture models
K-means, EM and Mixture models
Vu Pham
 
Speech recognition system seminar
Speech recognition system seminarSpeech recognition system seminar
Speech recognition system seminar
Diptimaya Sarangi
 

En vedette (20)

Text independent speaker recognition system
Text independent speaker recognition systemText independent speaker recognition system
Text independent speaker recognition system
 
iVector vs GMM/UBM for Automatic Speaker Recognition system
iVector vs GMM/UBM for Automatic Speaker Recognition system iVector vs GMM/UBM for Automatic Speaker Recognition system
iVector vs GMM/UBM for Automatic Speaker Recognition system
 
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
 
Speaker Recognition
Speaker RecognitionSpeaker Recognition
Speaker Recognition
 
Speaker recognition.
Speaker recognition.Speaker recognition.
Speaker recognition.
 
Indexation image
Indexation imageIndexation image
Indexation image
 
chap1 generalites_signaux-systemes
chap1 generalites_signaux-systemeschap1 generalites_signaux-systemes
chap1 generalites_signaux-systemes
 
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
Text Prompted Remote Speaker Authentication : Joint Speech and Speaker Recogn...
 
Presentation bpel
Presentation bpelPresentation bpel
Presentation bpel
 
Neurosciences et spiritualité fr_ Nancy -20131110
Neurosciences et spiritualité fr_ Nancy -20131110Neurosciences et spiritualité fr_ Nancy -20131110
Neurosciences et spiritualité fr_ Nancy -20131110
 
Feature Extraction
Feature ExtractionFeature Extraction
Feature Extraction
 
Expectation Maximization and Gaussian Mixture Models
Expectation Maximization and Gaussian Mixture ModelsExpectation Maximization and Gaussian Mixture Models
Expectation Maximization and Gaussian Mixture Models
 
Speaker Recognition using Gaussian Mixture Model
Speaker Recognition using Gaussian Mixture Model Speaker Recognition using Gaussian Mixture Model
Speaker Recognition using Gaussian Mixture Model
 
Speech recognition final presentation
Speech recognition final presentationSpeech recognition final presentation
Speech recognition final presentation
 
K-means, EM and Mixture models
K-means, EM and Mixture modelsK-means, EM and Mixture models
K-means, EM and Mixture models
 
Speech Recognition Technology
Speech Recognition TechnologySpeech Recognition Technology
Speech Recognition Technology
 
Speech recognition system seminar
Speech recognition system seminarSpeech recognition system seminar
Speech recognition system seminar
 
Artificial intelligence Speech recognition system
Artificial intelligence Speech recognition systemArtificial intelligence Speech recognition system
Artificial intelligence Speech recognition system
 
Speech Recognition System By Matlab
Speech Recognition System By MatlabSpeech Recognition System By Matlab
Speech Recognition System By Matlab
 
Speech recognition
Speech recognitionSpeech recognition
Speech recognition
 

Similaire à Irisa p gros

Content analytics slideshare aproged
Content analytics slideshare aprogedContent analytics slideshare aproged
Content analytics slideshare aproged
Aproged
 
Samar - Premier bilan d'étape - Oct. 2010
Samar - Premier bilan d'étape - Oct. 2010Samar - Premier bilan d'étape - Oct. 2010
Samar - Premier bilan d'étape - Oct. 2010
Stefane Fermigier
 

Similaire à Irisa p gros (20)

Metadonnees -- une typologie
Metadonnees -- une typologieMetadonnees -- une typologie
Metadonnees -- une typologie
 
Séminaire Ist inria 2014 : Pascale Sébillot
Séminaire Ist inria 2014 : Pascale SébillotSéminaire Ist inria 2014 : Pascale Sébillot
Séminaire Ist inria 2014 : Pascale Sébillot
 
Reconnaissance vocale et création artistique
Reconnaissance vocale et création artistiqueReconnaissance vocale et création artistique
Reconnaissance vocale et création artistique
 
Semantic Information Systems
Semantic Information SystemsSemantic Information Systems
Semantic Information Systems
 
Semantic Information Systems
Semantic Information SystemsSemantic Information Systems
Semantic Information Systems
 
Semantic Information Systems
Semantic Information SystemsSemantic Information Systems
Semantic Information Systems
 
L'empreinte audio numerique au service de l'analyse des diffusions Masterclas...
L'empreinte audio numerique au service de l'analyse des diffusions Masterclas...L'empreinte audio numerique au service de l'analyse des diffusions Masterclas...
L'empreinte audio numerique au service de l'analyse des diffusions Masterclas...
 
Soutenance.final
Soutenance.finalSoutenance.final
Soutenance.final
 
Arabic Speech Recognition System Using CMU-Sphinx4
Arabic Speech Recognition System Using CMU-Sphinx4Arabic Speech Recognition System Using CMU-Sphinx4
Arabic Speech Recognition System Using CMU-Sphinx4
 
Content analytics slideshare aproged
Content analytics slideshare aprogedContent analytics slideshare aproged
Content analytics slideshare aproged
 
Introduction à la fouille de textes et positionnement de l'offre logicielle
Introduction à la fouille de textes et positionnement de l'offre logicielleIntroduction à la fouille de textes et positionnement de l'offre logicielle
Introduction à la fouille de textes et positionnement de l'offre logicielle
 
Semantic Information Systems
Semantic Information SystemsSemantic Information Systems
Semantic Information Systems
 
Sem info system_2012
Sem info system_2012Sem info system_2012
Sem info system_2012
 
La voix avec common voice
La voix avec common voiceLa voix avec common voice
La voix avec common voice
 
Recursive Neural Network summary
Recursive Neural Network summaryRecursive Neural Network summary
Recursive Neural Network summary
 
les techniques TALN
les techniques TALNles techniques TALN
les techniques TALN
 
Normes standards (numériques)
Normes standards (numériques)Normes standards (numériques)
Normes standards (numériques)
 
Ecm Open Source
Ecm Open SourceEcm Open Source
Ecm Open Source
 
FRESNEL_Quentin_Rapport
FRESNEL_Quentin_RapportFRESNEL_Quentin_Rapport
FRESNEL_Quentin_Rapport
 
Samar - Premier bilan d'étape - Oct. 2010
Samar - Premier bilan d'étape - Oct. 2010Samar - Premier bilan d'étape - Oct. 2010
Samar - Premier bilan d'étape - Oct. 2010
 

Plus de Cédric WILLIAMSON

Plus de Cédric WILLIAMSON (11)

Hogunsoft presentation meito atelier crm
Hogunsoft  presentation meito atelier crmHogunsoft  presentation meito atelier crm
Hogunsoft presentation meito atelier crm
 
Hisseo presentation meito atelier crm
Hisseo presentation meito atelier crmHisseo presentation meito atelier crm
Hisseo presentation meito atelier crm
 
Hogunsoft presentation meito atelier crm
Hogunsoft  presentation meito atelier crmHogunsoft  presentation meito atelier crm
Hogunsoft presentation meito atelier crm
 
Logica presentation meito atelier crm
Logica presentation meito atelier crmLogica presentation meito atelier crm
Logica presentation meito atelier crm
 
Thales
ThalesThales
Thales
 
In pixal fusion_algos
In pixal fusion_algosIn pixal fusion_algos
In pixal fusion_algos
 
Advansee
AdvanseeAdvansee
Advansee
 
Jm Jezequel irisa Aom4 agility
Jm Jezequel irisa Aom4 agilityJm Jezequel irisa Aom4 agility
Jm Jezequel irisa Aom4 agility
 
L Morisseau Adoption De L Agilite
L Morisseau Adoption De L AgiliteL Morisseau Adoption De L Agilite
L Morisseau Adoption De L Agilite
 
Anteo Mda Aosd
Anteo Mda AosdAnteo Mda Aosd
Anteo Mda Aosd
 
Exibri Software Product Lines Aosd
Exibri Software Product Lines AosdExibri Software Product Lines Aosd
Exibri Software Product Lines Aosd
 

Irisa p gros

  • 1. Les enjeux scientifiques de l’indexation vidéo Patrick Gros Responsable de l’équipe TEXMEX INRIA Rennes et IRISA http://www.irisa.fr/texmex
  • 2.
  • 3.
  • 4.
  • 5.
  • 6.
  • 7.
  • 8.
  • 9.
  • 10.
  • 11.
  • 12.
  • 13.
  • 14.
  • 15.
  • 16.
  • 17.
  • 18.
  • 19. Web-based topic adaptation ... ... € … thus a candidate who fails to carry a particular state receives not a single electoral vote in that state for the popular votes received since residential elections are won by electoral ... candidate state election 3. Building of an adaptation corpus candidate vote electoral vote 2. Querying 1. Keyword spotting Adaptation LM 4.a Training of a topic-specific LM 4.b Mix of this LM and the general one Baseline LM + Adapted LM = Web search engine ✘ ✔ ✔ ✔ ✘ ✔ ✘ ✔
  • 20.
  • 21.