Analyse conjointe texte et image pour la caractérisation de films d'animation

Abstract : These last years, the fast development of new technologies allows digital media collections and the circulation of these data growing in size and number. However, the exploitation of these data remains a whole problem and creates a strong requirement for efficient tools to manipulate it. The current trend is in search of automatic indexing technique, based on semantic document contents. In the context of Annecy International Animation Film Festival the animated movie characterization presented in our works consists in information fusion between information contained in the animated movie images, textual information extracted from festival registration form and the expert knowledge. This information fusion uses color statistics and activity measure extracted from automatic image sequence analysis and uses textual description and emotion measure from automatic textual analysis. Two characterizations are provided from this information fusion. A first global characterization consists in a dramatic emotion classification and a second local characterization consists in the time-localized action description from the actant scenario.
Complete list of metadatas

Cited literature [188 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00750619
Contributor : Grégory Païs <>
Submitted on : Sunday, November 11, 2012 - 7:12:55 PM
Last modification on : Wednesday, January 10, 2018 - 9:44:43 AM
Long-term archiving on : Tuesday, February 12, 2013 - 3:45:21 AM

Identifiers

  • HAL Id : tel-00750619, version 1

Collections

Citation

Païs Grégory. Analyse conjointe texte et image pour la caractérisation de films d'animation. Intelligence artificielle [cs.AI]. Université de Savoie, 2010. Français. ⟨tel-00750619⟩

Share

Metrics

Record views

814

Files downloads

2335