Repérage d'images ordinaires : analyse des requêtes des chercheurs d'images.

For many years, the Web became an important media for the diffusion of multilingual resources. However, linguistic differences still form a major obstacle to scientific, cultural, and educational exchange. Besides this linguistic diversity, a multitude of databases and collections now contain documents in various formats, which may also adversely affect the retrieval process. For example, image collections have many areas of application. The need to retrieve images from a particular collection is shared by several user communities. Until recently, image collections were difficult to access due to limitations in dissemination and duplication procedures. The advent of the Web highlighted the pressing need to develop suitable tools for the description of digital images, since these can be found in the majority of available resources. This research compares image retrieval within two contexts: a monolingual context, i.e., where the language of the query (French) is the same as the indexing language (French); and a multilingual context, i.e., where the language of the query (French) is different from the indexing language (English).This article presents the results of the analysis of the queries used to retrieve a set of ordinary images representing everyday-life objects in a multilingual retrieval context. The analysis of the terms contained in the queries of the image searchers illustrates the terminological, perceptual, and structural trends. In general, participants employ simple queries containing only one word. Moreover, the queries are often generic in nature. The analysis also underlines that participants frequently use functional attributes to retrieve ordinary images. On the structural level, generic relations dominate other types of relations. The contributions of this research project are manifold. At the theoretical level, this research identifies the characteristics essential to the indexing process of digital images. This study also emphasizes the pressing necessity to optimize the methods used for image processing, in order to facilitate the image retrieval and their dissemination in multilingual environments. [ABSTRACT FROM AUTHOR]