Multiple-View Constrained Clustering For Unsupervised Face Identification In TV-Broadcast

Bendris, Meriem; Charlet, Delphine; Favre, Benoit; Damnati, Géraldine; Auguste, Rémi

Document type :

Communication dans un congrès avec actes

DOI :

10.1109/ICASSP.2014.6853645

Title :

Multiple-View Constrained Clustering For Unsupervised Face Identification In TV-Broadcast

Author(s) :

Bendris, Meriem [Auteur]
Laboratoire d'informatique Fondamentale de Marseille [LIF]
Charlet, Delphine [Auteur]
France Télécom Recherche & Développement [FT R&D]
Favre, Benoit [Auteur]
Traitement Automatique du Langage Ecrit et Parlé [TALEP]
Laboratoire d'informatique Fondamentale de Marseille - UMR 6166 [LIF]
Damnati, Géraldine [Auteur]
France Télécom Recherche et Développement [Lannion] [FTR&D]
Auguste, Rémi [Auteur]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Laboratoire d'Informatique Fondamentale de Lille [LIFL]
FOX MIIRE [LIFL]

Conference title :

ICASSP2014 - Image, Video, and Multidimensional Signal Processing (ICASSP2014 - IVMSP)

City :

Florence

Country :

Italie

Start date of the conference :

2014-05-04

Publication date :

2014

HAL domain(s) :

Informatique [cs]/Informatique et langage [cs.CL]

English abstract : [en]

Our goal is to automatically identify faces in TV broadcast without a pre-defined dictionary of identities. Most methods are based on identity detection (from OCR and ASR) and require a propagation strategy based on visual ...
Show more >Our goal is to automatically identify faces in TV broadcast without a pre-defined dictionary of identities. Most methods are based on identity detection (from OCR and ASR) and require a propagation strategy based on visual clustering. In TV content, people appear with many variations making the clustering difficult. In this case, speaker clustering can be a reliable link for face clustering. Multi-modal clustering methods assume a bipartite mapping between modalities. In this paper, we propose to build automatically an incomplete speaker-face mapping based on local evidence of OCR and Lip activity links. Then, we propose schemes of speaker constraints propagation to the face constrained-clustering problem. Experiments performed on the REPERE corpus show an improvement of face identification by propagating names to face clusters (+3.7% F-measure compared to the baseline).Show less >

Language :

Anglais

Peer reviewed article :

Oui

Audience :

Internationale

Popular science :

Non

Collections :