Nommage non-supervisé des personnes dans les émissions de télévision: une revue du potentiel de chaque modalité

Johann Poignant 1 Laurent Besacier 2 Georges Quénot 1
1 MRIM - Modélisation et Recherche d’Information Multimédia [Grenoble]
Inria - Institut National de Recherche en Informatique et en Automatique, LIG - Laboratoire d'Informatique de Grenoble
Abstract : ersons identification in TV broadcast is a valuable tool for indexing these videos. But the use of biometric models is an unsustainable option without a priori knowledge of people present in the videos. The names pronounced or written on the screen can provide us a list of hypotheses names. We propose a comparison of the potential of these two modalities (names pronounced or written) to extract the true names of the speakers and/or faces. The names pro- nounced offer many instance of citation but transcription and detection errors of these names halved the potential of this modality. The names written benefits of the video quality improve- ment and there are easy to find. The affiliation to speakers/faces of names written is simpler than for names pronounced.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [12 references]  Display  Hide  Download

https://hal.inria.fr/hal-00953087
Contributor : Marie-Christine Fauvet <>
Submitted on : Monday, March 3, 2014 - 4:36:03 PM
Last modification on : Friday, March 22, 2019 - 12:56:02 PM
Document(s) archivé(s) le : Saturday, May 31, 2014 - 10:45:31 AM

File

POIGNANT--CORIA--2013.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00953087, version 1

Citation

Johann Poignant, Laurent Besacier, Georges Quénot. Nommage non-supervisé des personnes dans les émissions de télévision: une revue du potentiel de chaque modalité. CORIA 2013, papier long (oral), 2013, Neuchatel, Suisse. 2013. 〈hal-00953087〉

Share

Metrics

Record views

231

Files downloads

128