Organizing moving image collections for the digital era : the goldspeil report

Bibliographic reference
M. TURNER James, HUDON Michele, DEVIN Yves. Organizing moving image collections for the digital era : the goldspeil report". Information Outlook. August 2002. FindArticles.com. 15 Feb. 2007.

Text accessible : http://www.findarticles.com/p/articles/mi_m0FWE/is_8_6/ai_90932749

DC Title : Organizing moving image collections for the digital era: the goldspeil report
D.C Creator : M. TURNER James, HUDON Michele, DEVIN Yvest
D.C Subject : digital image/images collections/moving image/organization/research/retrieval/storage/support
D.C Description : this text explains the goal which is to understand the techniques and tools used for representing the content of moving image collections that are indexed shot by shot.
D.C Publisher : nformation outlook
D.C Date : August 2002
D.C Type : Text
D.C Format :html.
D.C Identifier : http://www.findarticles.com/p/articles/mi_m0FWE/is_8_6/ai_90932749
D.C Source : http://www.findarticles.com/
D.C Language : english
D.C Relation : -
D.C Coverage : -
D.C Rights :© COPYRIGHT 2002 Special Libraries Association/Copyright 2003 Gale Group



This text is an extract of the original text.


Organizing moving image collections for the digital era : the goldspeil report

Information Outlook, August, 2002 by James M. Turner, Michele Hudon, Yves Devin

"THE STEVEN I. GOLDSPIEL MEMORIAL RESEARCH FUND WAS ESTABLISHED IN 1991, and named in honor of the former president of Disclosure, Inc. (now known as Primark). The research fund is an endowment and projects are funded solely from investment income generated by the fund. The purpose of the fund is to support projects that promote research and advancement of library sciences, in particular focusing on projects that address the goals identified in the SLA Research Statement.

The 1999 winners of the award were James M. Turner, Michele Hudon and Yves Devin from the Universite de Montreal. The results of their project titled, "Organizing Moving Image Collections for the Digital Era" are presented here.

Introduction

Pictures have always been used to represent concepts and ideas and to communicate messages. Now that we collect them so extensively, we need to represent the pictures themselves in order to store and retrieve them. Photography, movies, television and digital images stored on computers have all contributed to the rapid buildup of ever larger collections. Whatever the format or the presentation medium, pictures have become a most important mode of communication in our time. They play a crucial role in such areas as medicine, journalism, advertising, education and entertainment. The notion of picture collections now has to do with a vast world, and an attempt to describe this world is represented by the study poster entitled The World of Visual Collections (GRIV 1998), which takes into account the areas of art, engraving, photography, computer graphics, the types of institutions which collect and the many professions that use pictures as part of their work.

Moving images are, of course, a goldmine for many organizations and individuals, and it is important to describe them adequately in databases in order to show their richness and complexity if we are to exploit these collections fully. However, the world of moving image collection organization is one of locally established techniques, with little or no standardization and without communicability between systems. This has not been a problem until now because systems were managed independently of each other, but in the networked world in which we now live, the question of discovery and exchange of information has come to the forefront and needs to be addressed.

Our recently-completed research project was concerned with indexing moving image collections for storage and retrieval. The general goal of the project was to understand the techniques and tools used for representing the content of moving image collections that are indexed shot by shot. We especially wanted to study the question of indexing languages and their structure, as well as techniques for keeping them current. Several more specific objectives were identified:

* to determine how many terms, excluding proper names, are used to describe North American moving image collections indexed at the shot level;

* to estimate the rate of growth of term creation in these tools;

* to discover to what degree the lexical concepts are similar among the various tools; and

* to evaluate the possibility of creating a universal indexing vocabulary for general collections of moving images, those that represent everyday objects and events.

In this article we look at the background information to the study, after which we describe the method used to collect the data. The results are then given along with discussion of them, followed by conclusions we might draw from this research.

Background

Both still and moving images can be divided into three broad categories: art images, documentary images and "ordinary" images, each requiring its own type of organization. The proliferation of supports and the changing context (brought about first by the arrival of computer technology and then the networking of resources) are the driving forces behind a great deal of work in retooling and rethinking work methods, but they do not change this fundamental (if arbitrary) classification. Nor does the shift from analogue to digital images. Thus, there is a great deal of work to do, but the guiding principles remain the same.

The many new systems and ever more efficient technologies for capturing and processing moving images require the establishment of effective management systems. It is necessary to be able to find any specific shot in a particular collection rapidly and efficiently. Without the establishment of new methods for storage and retrieval of moving images, these valuable resources will get lost in a hopeless jumble of useless data.

Research in the area of storage and retrieval of moving images takes place using two distinct approaches with little in common (Cawkell 1992, 180). The low-level or content-based approach is the focus of work by computer science researchers. This approach involves the statistical manipulation of pixels to get information about color preponderance and arrangement, recognition of textures, patterns, boundaries, objects, scene detection and so on. The high-level or concept-based approach is the focus of work by information science researchers. This approach involves human generation of metadata substantially assisted by computer technology (semi-automatic), as well as automatic generation of high-level metadata. The general focus of this approach is finding ways to generate shot-level indexing automatically from text created during the pre-production, production, and post-production stages, such as closed captioning, audio description, and production scripts. The two research streams are complementary, and the b est information systems for storage and retrieval of moving images will need to incorporate both approaches."[...]

Text as a Tool for Organizing Moving Image Collections

Bibliographic reference.
M. TURNER James, HUDON Michèle, DEVIN Yves, Text as a Tool for Organizing Moving Image Collections (en ligne), CAIS 2000 (Canadian Association for Information Science) : Dimensions of a
global information science, Proceedings of the 28th Annual Conference, 7 p.
Text accessible : http://www.slis.ualberta.ca/cais2000/turner.htm


D.C Title : Text as a Tool for Organizing Moving Image Collections
D.C Creator : TURNER M. James, HUDON Michèle, DEVIN Yves
D.C Subject : analysis/image collections/indexing/moving images/thesaurus

D.C Description : this text develops "common methods for shot-level and scene-level description of moving image documents, in order to foster discovery and retrieval of these resources worldwide."
D.C Publisher : CAIS 2000 (Canadian Association for Information Science) : Dimensions of a global information science, Proceedings of the 28th Annual Conference. Editor: Angela Kublik
D.C Date : -
D.C Type : Text.
D.C Format : html.
D.C Identifier : http://www.slis.ualberta.ca/cais2000/turner.htm
D.C Source : http://www.slis.ualberta.ca/
D.C Language : English
D.C Relation : Actes du Colloque http://www.slis.ualberta.ca/cais2000/index.htm
D.C Coverage : Canada
D.C Rights : -


This text is an extract of the original text.

Text as a Tool for Organizing Moving Image Collections
James M. Turner, Michèle Hudon, and Yves Devin
Université de Montréal

"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.

1. Introduction
As we enter the new millenium, we observe that the organization of moving image collections is still characterized by ad hoc information systems. In the rapidly-growing networked environment, it is critical to develop common methods of shot-level and scene-level description, in order to foster retrieval and ultimately, to share resources worldwide. Researchers are aware of the problem, and two streams of research which complement each other address the issues involved. The first stream focuses on low-level access to images using methods from computer science and concentrates on statistical techniques for deriving characteristics of images that help promote retrieval. The second stream focuses on high-level access to images using methods from library and information science, and concentrates on the use of text to create information useful for retrieval, information which is especially valuable since it is not available from the images themselves.
The materials contained in film libraries, television libraries, and both film and television stockshot libraries is usually of a general nature. Thus managing the terminology used to index these collections involves descriptor lists and thesauri which are often constructed from scratch to reflect the particular local reality of the collection. For the researcher who looks in many collections to find material for use in film and video production, this means relying heavily on the resource persons who work with each collection and learning a number of different retrieval systems and indexing vocabularies.
It has been thought that a thesaurus for indexing everyday film and video materials would become unmanageable after a certain point because there are so many kinds of persons, objects and events to describe that it would eventually become impossible to manage the semantic relationships between them. However, anecdotal evidence suggests that term creation levels off once sufficient terminology for indexing most shots has been created. This paper describes a research project which belongs to the second or high-level research stream and which aims to study the characteristics of thesauri used for indexing moving image collections at the shot level.
What is the point at which term creation levels off? How many terms for describing moving images does a thesaurus need to contain in order to be considered complete enough to describe a general collection adequately? Are the terms similar from one thesaurus to the next, or are collections so particular that an individual tool is required for each collection? Would it be reasonable to try to construct a general thesaurus of everyday persons, objects and events that could be shared among moving image collection managers? These are the research questions our project addresses.
The general goal of the study is to reach an understanding of the organization of existing vocabulary-management tools for moving image collections. The specific objectives are:
to discover how many terms, excluding proper names, are contained in a controlled vocabulary for managing general moving images collections before term creation levels off;
to identify patterns among terms in the existing thesauri created for moving image collections;
to assess how patterns found can contribute to building a shared vocabulary useful for special collections containing general material.

2. Background
Moving image collections are largely found in movie and television production facilities, but they are also to be found in many other contexts and environments, such as corporate libraries, government agencies, documentation centres of research groups, holocaust museums, religious archives, and so on. Librarians have developed a great deal of expertise with managing print collections, but there are no generally- and widely-accepted tools available for organizing moving image collections. Visual resources librarians and researchers are working on problems related to art collections and to slide collections of works of art, but moving image and other non-art picture collections need urgent attention. This is due to the profusion of production and the consequent mushrooming of such collections in recent decades, including television news libraries and stockshot libraries. To be of help to their user base, these collections need to be catalogued and indexed at the shot level.
In the context of working groups around the planet trying to work out metadata standards for the management of all kinds of digital materials, high-level metadata standards for moving image databases have not been studied. We hope that the results of our study will make a contribution in helping establish a theoretical basis on which standards can be built." [...]