Content uploaded by Michèle Hudon
Author content
All content in this area was uploaded by Michèle Hudon on Sep 30, 2015
Content may be subject to copyright.
A preview of the PDF is not available
Text as a Tool for Organizing Moving Image Collections
Abstract
In the rapidly-growing networked environment, it is critical to develop common methods for shot-level and scene-level description of moving image documents, in order to foster discovery and retrieval of these resources worldwide. In this paper, we describe the methodology used and report preliminary results obtained in a project designed to study existing tools for indexing moving images at the shot level. A number of institutions holding moving image collections were recruited as partners, and visits to them were then made to complete a questionnaire and gather other information during a structured interview. Preliminary results show that a variety of tools are used for indexing moving images, and that those institutions which have been able to invest in building good indexing tools over the years have seen the quality of those tools suffer in the wake of cutbacks in resources needed to maintain them.
... Ce n'est pas le cas actuellement. Et enfin, on peut penser à un problème générique lié à l'utilisation d'un outil ad hoc comme totem, qui est la limitation en termes d'échange de données et d'interopérabilité entre différents centres documentaires, ou pour mise à disposition de contenus sur Internet (voir [Turner et al., 2000]). En résumé, on peut dégager les besoins suivants dans la tâche documentaire : représenter le contenu sémantique d'un document de manière non ambigüe et de manière à ce que cette représentation soit opérationnelle pour une gestion informatique du système (nécessaire dans le cas de grosses bases documentaires). ...
Resources like terminologies or ontologies are used in a number of applications, including documentary description and information retrieval. Different methodologies have been proposed to build such resources, on the basis of experts' interviews or of textual corpora.
This thesis focuses on the use of existing Natural Language Processing methodologies, meant to help the building of ontologies from textual corpora, to build a particular type of resource : differential ontologies. These ontologies are structured according to a system of semantic identities and differences between their constituents: terms of the domain and categorisation items called “top level categories”.
We present different experiments that we have done to elicit, structure, define and “interdefine” the terminological items relevant for a given task. Our first use case was the opales pro ject, in which we had to provide a group of anthropologists with the conceptual vocabulary that they needed to annotate audiovisual documents about childhood. We have used the textual corpus that we have built in this pro ject to test linguistic tools
and methodologies for building ontologies from textual data, and we have defined our own programs. The suite of resulting programs is called SODA, and they focus on the extraction and use of defining contexts in corpora to spot terminological items, to structure them and to provide semantic similarity information that enables to compare them.
Cet article présente l’analyse de fiches descriptives de banques d’images et leur normalisation. Il présente tout d’abord quelques concepts théoriques sur les images et leur description, sur les modes de recherche et de repérage des images, sur les métadonnées internes (EXIF, IPTC et XMP) ainsi que sur la normalisation. L’analyse des éléments de la description formelle des photographies dans 30 banques d’images tient compte des éléments potentiellement utiles pour la recherche et le repérage, de la terminologie employée, de la longueur des légendes et de la fréquence d’utilisation des métadonnées internes. L’article conclut que les fiches descriptives ne sont pas normalisées (bien qu’il existe une tendance à réutiliser quelques éléments coïncidant avec les champs des métadonnées IPTC) et recommande d’employer des métadonnées internes pour normaliser la description formelle des images.
ResearchGate has not been able to resolve any references for this publication.