Cataloging/Metadata
Visual Image Retrieval: an image is worth a thousand words.
The topic of this research paper has been debated since the 1980’s by groups of professionals working in disciplines with different objectives. Since the field is experiencing a very rapid and multi-directional growth, the document is particularly apt at synthesizing the different views and evaluating their weight with a comprehensive approach.
Abstract
A networked environment has generated numerous demands on all aspects of information access and hence on metadata schema proponents, metadata creators, tool providers and educational organizations (service providers.) Visual images about any topic are a particular challenge—when we strive to represent or describe them. The difficulty of describing an image is nothing new since whoever assigns descriptors or subjects to a resource, even if from a controlled vocabulary, becomes an author and as such separates herself from the readers (the system users). What are the user’s expectations about the workings of retrieval systems, we may ask? As ever larger numbers of individuals gain the means to access online information or the services, such as libraries, that can offer that access for free, the demand for easy to use retrieval system grows louder. The visual image collections are no exception to this trend, on the contrary the social shift to multimedia communication is increasing their center stage position. Can today’s metadata structures be (or become) the container for the data that would facilitate the indexing of the visual image story, characters, events and spaces as well as mood, emotion, and theme? This study of visual image searches confirms the need of relying on a diverse set of metadata and that no one metadata schema may be sufficient to satisfy the access needs of the non-expert user. Instead a combination of techniques needs to be employed and than merged using an easy to use front-end interface. This interface has to provide the user with a clear exposition of available materials and, possibly, options to reuse the information as the user sees fit.
Kaleidoscope: an overview of metadata research
The library may embody three primaries roles: it is the locus of information-found, it is a place for education exchange (informal learning) and it is an agorà where the civic and the recreational intersect (outreach and planned events.) The interest of this writing focuses on libraries’ role as locus of information-found. As such loci, libraries are nodes in an information universe in which information is concentrated and separated from the surrounding vacuum and noise. This information is imagined to be neatly contained in “information packages.” The concentration of information is function of the scope of those in need to find it as long as said information is easily accessible (retrievable.) To that end, librarians devised systems to describe information and organize it. Four factors seem to have weighed on the development and implementation of these information organization systems: the designers’ mental models about which elements are relevant to the description of the information package; the interpretation of the information package content; the physical nature of the information package; and the assumptions about the anticipated user—or their lack thereof. The continuous development of the form and type of information packages presents a great challenge to the capability of any static model to describe them and promotes the research on ever new information description models—metadata schema. The impetus for the research is fueled by the implicit promise that the possibilities of a networked environment presents—unfettered accessibility.Whereas textual documents offer a variety of entries to their description in a textual environment, non-textual resources present numerous challenges as the description (the aboutness) will rely, in greater part, on the individual metadata creator’s visual perceptions, her mental models and/or the granularity of the metadata schema employed. It is this paper intention’s to lay the ground work for an understanding of the issues and possible solutions that creating searchable online repositories of visual materials entail and to frame this understanding within the contest of the discourse about metadata.