Metadata for images

A report for the FILTER project

Document details

Summary

This document provides a state-of-the-art report on image metadata for the FILTER (Focusing Images for Learning and Teaching - an Enriched Resource). After a brief introduction to image retrieval and use, the report describes the main categories of metadata that images require. There follows some short descriptions of particular metadata initiatives. These include: the Dublin Core Metadata Initiative, MPEG-7, the AMICO Data Dictionary, the SCRAN data standard, SPECTRUM, METS, the NISO draft Technical metadata for digital still images, the FGDC Content Standard for Digital Geospatial Metadata and ISO 19115. The following section describes some of the most prominent educational metadata initiatives, including the IEEE Draft Standard for Learning Object Metadata, the IMS Meta-data Specification and the proposal of the DCMI Education Working Group. A final section looks at some subject schemes that have specifically been designed for image data, mostly for art and architecture (e.g., AAT, and ICONCLASS) and for image collections in libraries (e.g., TGM I).

1. Introduction

1.1 Image retrieval and use

From prehistoric times, human communication has depended upon the creation and use of image-based information. Images have been a key component of human progress, for example, in the visual arts, architecture and geography. According to Ferguson (1992, p. 75), technological developments relating to images in the Renaissance, including the invention of printing and the use of linear perspective, had a positive effect upon the rise of modern science and engineering. The invention of photography and moving-image technology in the form of film, television and video recordings has increased global dependence on communication through images.

The importance of this image material is such that many different types of organisation exist in order to create, collect, maintain and give access to collections of them. These include organisations that are at least partly funded from public expenditure, e.g. art galleries, museums and libraries, as well as commercial organisations like newspapers and television companies (Evans, 1992). The types of image included in these collections can be extremely diverse and might include things like paintings, engineering diagrams, photographic prints, maps and films. Traditionally these organisations would have kept these images in their original formats (some of which will be obsolete). However, an increasing amount of image-based data is now becoming available in digital form, both through the success of digitisation initiatives and the growing use of digital technology in the image creation process (e.g. remote sensing or the use of digital cameras). This growth has led to a reappraisal of image retrieval techniques.

In general, there are two complimentary approaches to image retrieval, often described as 'content-based' and 'concept-based' (e.g., Rasmussen, 1997, p. 170). Content-based image retrieval (CBIR) refers to systems that can identify images by colour, shape or texture (e.g., Gudivada & Raghavan, 1995; Eakins & Graham, 1999). Content-based retrieval is a popular research topic in computing science, and potentially will have significant applications in biomedicine, trademark protection, Web-content filtering and other areas. CBIR systems that have been developed include IBM's Query By Image Content (QBIC) system (Flickner, et al., 1995) and experimental Web-based systems like Informedia (Wactlar, et al., 1996) and WebSEEk (Smith & Chang, 1997; Chang, et al., 1997). Despite the significant amount of research effort that is being devoted to CBIR, there is, however, a perception that the current generation of content-based systems cannot solve all image retrieval problems by themselves. For example, Chen, et al. (2001, p 831) make the point that "low-level visual features cannot represent the high-level semantic content of images." For this reason, there frequently remains a need for descriptive information (metadata) about images that can help facilitate retrieval at the semantic level. This is often known as concept-based retrieval.

Concept-based retrieval is based on the interpretation and description of images in terms of what they are and what they represent (Rasmussen, 1997, p. 170). There are two main ways in which this interpretation can be recorded: in a free-text 'caption' and by the assignment of one or more index terms from a controlled vocabulary. Image search and retrieval systems are then able to search both the controlled index terms and the captions associated with images. As with text-based information, however, concept-based image retrieval raises problems of subjectivity and 'aboutness.' Theoretical approaches to concept-based retrieval often refer to distinctions between levels of meaning made by the art historian Panofsky with regard to Renaissance art. Panofsky (1955, pp. 28-41) distinguished between three levels of meaning:

• Primary or natural subject matter, factual and expressional - pure forms, objects, events, expressions, etc. These motifs could be enumerated in a pre-iconographical description.
• Secondary or conventional subject matter - about representation, the connection of motifs with themes or concepts. Iconographical analysis deals with images, stories and allegories and presupposes that the interpreter is familiar with certain themes or concepts. One of Panofsky's examples being a particular arrangement of figures seated at a dinner table being a representation of the Last Supper.
• Intrinsic meaning or content - deals with 'symbolical' values. Iconological interpretation [1] helps to ascertain the "underlying principles that reveal the basic attitude of a nation, a period, a class, a religious or philosophical persuasion - qualified by one personality and condensed into one work" (Panofsky, 1955, p. 30).

Enser gives an example of this type of analysis based on Salvador Dali's painting "Christ of St John of the Cross." Enser (1995, p. 141) notes that this "displays pre-iconographic features (male figure, wooden cross, nails through palms of hands affixing figure to cross, wooden sailing ship, etc.), iconographic features (St. John, Crucifixion) and iconologic features (religious persecution, suffering, etc.)."

Shatford (1986) has further developed Panofsky's hierarchy of meaning in the context of analysing the subject of pictures for retrieval. For both Panofsky's primary and secondary levels, she distinguishes between what a picture is (objectively) of and what it is (more subjectively) about.

At the pre-iconographic level, the Of aspect is a generic description of objects and events; at the iconographic level, it is a specific, or proper, appellation of those objects and events. Of words describe people, places, objects, conditions, and actions that have a physical manifestation. The About aspect is, at the pre-iconographic level, a description of the mood of the picture; at the iconographic level the About aspect is an identification of mythical beings that have no unique and verifiable concrete reality, of symbolic meanings and abstract concepts that are communicated by images in the picture (Shatford, 1986, p. 45).

She also proposes a classification scheme in which subjects can be characterised as Generic Of, Specific Of, and About, with facets answering the questions: Who? What? When? and Where? (Rasumssen, 1997, p. 178).

Concept-based retrieval is usually achieved by the application of terms from a controlled vocabulary (or vocabularies) and a free-text description of an image, a title or caption. In addition, information can be recorded about other image attributes like creators or time and place of creation (Constantopoulos & Doerr, 1995). This means that concept-based image retrieval is, in the broadest sense, based on the creation and maintenance of metadata. Where reliable metadata is not available, e.g. the Web, some image search systems look for associated text (e.g. HTML ALT tags, adjacent headers, filenames, etc.) in an attempt to extract information about the semantics of images (Chen, et al., 2001; Goodrum & Spink, 2001).

After some introductory sections, this report will introduce a number of metadata-based initiatives that have relevance to image retrieval. Some of these will be initiatives based in the cultural heritage domain, e.g. AMICO or SCRAN. Others will be the product of digitisation initiatives and will have a mainly technical focus on metadata, e.g. METS or the NISO draft Technical Metadata for Digital Still Images. Further sections will look at geospatial and educational metadata initiatives.

1.2 The FILTER project context

The FILTER (Focusing Images for Learning and Teaching - an Enriched Resource) project is investigating the use of digital images as learning and teaching resources. The project will be setting up an exemplar database of images and learning resources that will be supplied by Learning and Teaching Support Network (LTSN) Subject Centres. This report is intended to provide an evaluation of the current state of image metadata developments and feed into the development of the two metadata element sets that will be used in the exemplar database, one for the images stored in the database, the other for the learning resources supplied by the LTSN centres.

1.3 Metadata and its uses

Metadata is usually defined as 'data about data' but normally refers to structured machine-understandable data about data. In the library community, metadata is often used to refer to the type of descriptive information contained in catalogue records; bibliographic-type information that describes resources together with selected index headings. Metadata, therefore, acts as a basis for information retrieval but can also have other functions, e.g. the management of user-access to resources, or preservation.

The diversity of metadata creating communities, however, results in the existence of many different metadata formats. Different subject communities and market sectors have invested heavily in developing their own metadata formats and systems. Dempsey & Heery (1998) have pointed out that considerable effort has been expended on developing specialist formats to ensure fitness for purpose.

... there has been investment in training and documentation to spread knowledge of the format; and, not least, systems have been developed to manipulate and provide services based on these formats.

For these reasons, metadata 'format diversity' is likely to be perpetuated over time and, indeed, new metadata formats will periodically be developed to address the perceived needs of other subject domains and communities.

Historically, different types of repositories have developed their own ways of describing the content of their collections. Libraries have developed standards for describing library-based materials including images, using a variety of standards like the International Standard Bibliographic Description (ISBD) series, the Anglo-American Cataloguing Rules and formats like MARC21. Archives use a different family of standards including ISAD(G), the General International Standard Archival Description. Museums and art galleries, if anything, are even more diverse, although mda's SPECTRUM is a useful attempt to develop a unified standard based on the practice of many museum professionals. In general, however, individual museum departments often document their holdings in a number of different ways, depending upon the provenance and format of items. Blackaby and Sandore (1997) have commented that for "a variety of reasons - availability of software, variations in standards, the needs of collections - heterogeneity can be expected to persist and in many cases, it should be encouraged." When other organisations that hold image collections are included - e.g., photographic libraries, local history societies, healthcare organisations (e.g., Wong & Tjandra, 1999), scientific data centres (e.g., Carazo & Stelzer, 1999; Gonzalez-Couto, Hayes & Danckaert, 2001), etc. - one gets an idea of the potential heterogeneity of image-based data and its associated metadata. This diversity will also be replicated in the digital environment. User expectations, however, mean that there is a need to work towards integrating the resource discovery process in this heterogeneous environment.

The creation and maintenance of metadata will be an important part of the activity of any image archive - whether digital or non-digital. Image repositories like photographic libraries have traditionally used cataloguing techniques to store metadata about their resources. These will often, when known, record an object's creator and date of creation and may also contain information on its physical form and subject matter (Constantopoulos & Doerr, 1997). Images also have their own subject indexing requirements so that some classification schemes have been developed specifically for image-based resources. One of the best known of these is ICONCLASS, a hierarchical classification system for classifying works of art (Couprie, 1983). Text-based indexing schemes include the widely used Art & Architecture Thesaurus (AAT) and the Categories for the Description of Works of Art (CDWA), both of which are products of the Getty Vocabulary Program. Some relevant subject schemes are described in more detail in section 7.

1.4 Image metadata

Howard Besser of the University of Michigan has proposed several distinct areas where metadata standards need to be developed for digital images. He has said that the library and information community - amongst others - need to take "the steps necessary to ensure that digital images produced today will be viewable well into the future, and a key step in making that happen is the provision of adequate metadata" (Besser, 1997). This metadata comes in six broad categories (Besser and Trant, 1994).

• The technical information required to view the image. This metadata would comprise information about image types - whether bitmaps, vector files or video - with information on particular file formats (e.g. TIFF, GIF, etc.), compression (e.g. JPEG) and colour metrics.
• Information about the image capture process. This would include metadata about the type of image digitised with information about the size and dimensions of the original object. Technical information about the maker and model of scanner hardware used together with details of what has been done to each image would also be useful.
• Information about the quality and veracity of an image. Users of images may need to know who was responsible for its digitisation. For example, a digital image created by a museum or art gallery from scanning the original or a high-quality surrogate may need to be differentiated from the same image scanned by a private individual at home.
• Information about the original object. It will in many cases be important to know the precise nature and origin of the source object and link it with any surrogates. This descriptive information may include legacy content metadata, including subject classifications.
• Information about an image's authenticity. One of the problems with all digital information is that it is difficult to be sure that an information object is what it claims to be. This is what Graham (1994) calls 'intellectual preservation'. How will users know that the digital object that they retrieve is the one that they want? How will administrators of digital repositories know that their holdings have not been subject to unauthorised changes, either accidental or deliberate? Solutions to these problems are likely to depend upon cryptographic techniques or implementations of digital signatures.
• Information about rights management. Rights information is basic to the use and reuse of image resources. Rights metadata might include information on viewing and reproduction restrictions and contact information for the rights holders. Publishers and other rights owners are beginning to investigate the potential of metadata for rights management (Rust, 1998).

Besser suggests that standards need to be developed for all of these metadata types. Decisions also need to be made concerning the location of this metadata. Some information, for example core resource discovery data or rights management metadata, might best exist in an image header while other information might more suitably be stored in a separate (but linked) database. It is essential that much of this information is recorded (or captured) at the time when a digital image is created.

2. General metadata initiatives

2.1 The Dublin Core Metadata Initiative (DCMI)

The Internet has spurred increased consideration of metadata as an aid to resource discovery. The Dublin Core Metadata Initiative (DCMI) is an international and interdisciplinary initiative to develop a metadata element set intended to facilitate the discovery of digital resources. Dublin Core was initially conceived as a simple metadata format that could be used by the creators of resource or by Web-site maintainers. It has also, however, become a focus of interest from a variety of communities with wider interests in resource description, including librarians, archivists, museum documentation specialists and computer scientists with an interest in text markup issues (e.g., Weibel & Lagoze, 1997).

International representatives of these communities have met at a series of invitational workshops, the first of which met at OCLC's headquarters at Dublin, Ohio in March 1995 which resulted in the initial proposal of an thirteen element metadata set (Weibel, et al., 1995). Dublin Core elements are intended to be optional, repeatable and extensible. The requirement for extensibility was recognised from an early stage because local applications of Dublin Core would need to add additional elements according to special needs (Weibel, 1999). The third Dublin Core Workshop specifically considered the application of the Dublin Core element set to the description of images. Dublin Core had originally developed its scope to "document-like-objects" - as a means of side-stepping what Weibel and Miller (1997) called "differences on individual notions of what constitutes a discrete object worthy of separate description." The workshop concluded that discrete and bounded images could be considered to be document-like-objects when 'fixed' - in the sense of appearing the same to all users. The workshop concluded that Dublin Core could constitute a suitable basis for the resource discovery of networked images. In addition, the workshop noted the benefits of adopting a common set of metadata elements that would support the discovery of both textual and visual resources (Weibel and Lagoze, 1997). The workshop resulted in a number of changes to the Dublin Core element set which took into account the specific requirements of images. Some of the element descriptions were changed and two new elements were added to the original thirteen. Thus the Subject and Description elements were differentiated and a new Rights Management element added. These fifteen elements (Table 1) are currently a DCMI Recommendation (DCMI, 1999).

Table 1: Dublin Core Metadata Element Set, v. 1.1

Source: http://dublincore.org/documents/1999/07/02/dces/

The fifteen Dublin Core elements can be refined by the use of qualifiers. These fall into two main categories: those qualifiers that make the meaning of an element more specific (Element Refinements) and those that identify schemes that aid the interpretation of an element value (Encoding Schemes). An example of an Element Refinement would be the addition of an 'abstract' qualifier to the 'Description' element. An example of an Encoding Scheme would be the indication of the use of a controlled vocabulary in the 'Subject' element, or the use of a particular form of 'Date,' e.g. the W3C Encoding rules for dates and times. The principles governing the use of Dublin Core qualifiers and those approved by the Dublin Core Usage Committee have been published by DCMI (2000).

The Dublin Core Metadata Element Set (DCMES) is used by a number of initiatives that describe images. For example, the participants in PictureAustralia (http://www.pictureaustralia.org/) use the Dublin Core to describe digitised images that illustrate the cultural heritage of Australia. The National Library of Australia hosts a central metadata index that is based on DCMES. Each participant provides a slightly different 'flavour' of DC, as the data is mapped from a variety of legacy systems (http://www.pictureaustralia.org/metadata.html). Participants in the initiative have also agreed to use where possible the Australian Pictorial Thesaurus (APT), a hierarchical thesaurus of 15,000 subject terms (http://www.picturethesaurus.gov.au/). The Scottish Cultural Resources Access Network (SCRAN) also uses an implementation of the Dublin Core to describe images.

More information on DCMI can be found at: http://dublincore.org/

2.2 Multimedia Content Description Interface (MPEG-7)

MPEG-7, the Multimedia Content Description Interface , is part of a family of standards developed by the Moving Picture Experts Group (MPEG) - a working group (JTC1/SC29/WG11) of the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC). MPEG (http://mpeg.telecomitalialab.com/) are responsible for the development of standards for the coded representation of digital video and audio. While other MPEG standards (e.g., MPEG-1, MPEG-2 and MPEG-4) have focused on the coded representation of audio-visual content, MPEG-7 is primarily concerned with information about content, i.e., metadata.

Development of MPEG-7 began in October 1996. It was motivated by the need to extract or manipulate information from large collections of multimedia content, e.g. the outputs of video surveillance or Internet live feeds (Koenen & Pereira, 2000, p. 6). The standard is intended for use in a variety of applications, including the broadcast media, digital libraries, e-commerce and multimedia directory services, and covers still images, graphics, three dimensional models, audio, speech, video, and combinations of them. Nack & Lindsay (1999a, p. 70) have said that the ultimate aim of MPEG-7 was to make audio-visual material as searchable as text.

The basic concepts used in MPEG-7 are 'Descriptors' (that define the syntax and the semantics of each feature or metadata element) and 'Description Schemes' that specify the structure and semantics of the relationships between components. In addition, MPEG-7 specifies a 'Description Definition Language' (DDL) that provides a syntax for the express, combine, extend and refine Descriptors and Description Schemes (Hunter & Nack, 2000; Hunter, 2001). This is based on the XML Schema language (http://www.w3.org/XML/Schema). MPEG-7 Descriptors are intended to apply at different levels of abstraction, e.g. from low-level visual features like shape, size, texture and colour to high level 'semantic' information, e.g. about abstract concepts, events, genres, etc. (Chang, Sikora & Puri, 2001, p. 689). These can then be associated with other metadata (Descriptors and Descriptor Schemes) about the coding schemes used (e.g. MPEG-2), conditions of access, rating information, context, etc.

The MPEG-7 specification is formally known as ISO/IEC 15938 (Information technology -- Multimedia content description interface) and is organised into seven parts:

• ISO/IEC 15938-1: Systems
• ISO/IEC 15938-2: Description Definition Language
• ISO/IEC 15938-3: Visual
• ISO/IEC 15938-4: Audio
• ISO/IEC 15938-5: Multimedia Description Schemes
• ISO/IEC 15938-6: Reference Software
• ISO/IEC 15938-7: Conformance testing

MPEG-7 Part 3 specifies a set of standardised low-level Descriptors and Description Schemes for visual content, e.g., the shape of objects, object size, colour, texture, motion, etc. (Sikora, 2001).

MPEG are now working on a new standard, the MPEG-21 Multimedia Framework (to become ISO/IEC 21000) that attempts to provide a framework for the integration of standards to support the delivery and consumption of multimedia content (Bormans & Hill, 2001). MPEG-21 includes specifications for a Digital Item Identification and Description (MPEG-21 Part 3) that will uniquely identify digital content. Further specifications of a Rights Expression Language and Data Dictionary (MPEG-21 Parts 5 and 6) are intended to facilitate the interchangeable expression of intellectual property rights information (Koenen, 2001).

General information on MPEG 7 can be found in: Nack & Lindsay (1999a; 1999b), Koenen & Pereira (2000), Chang, Sikora & Puri (2001) and at: http://ipsi.fhg.de/delite/Projects/MPEG7/

3. Cultural heritage initiatives

The cultural heritage domain has been very active in metadata activities. Much of this activity has not been directly been focused on the description of images, but there are significant initiatives with relevance to images that have mainly originated in the museum sector. Examples of these might be the Museum Educational Site Licensing Project (MESL), the Art Museum Image Consortium (AMICO) and SCRAN, all of which were concerned with making digitised versions of artworks and other objects available for educational use. Other initiatives are more focused on system design or interoperability, e.g. the projects Aquarelle, COVAX (Contemporary Culture Visual Archives in XML),[2] ELISE (Electronic Library Image Service for Europe) and the SPECTRUM-XML development.

3.1 The Aquarelle project

The Aquarelle project, funded in the late 1990s by the European Union (EU) under its Telematics Applications Programme, investigated technical and architectural solutions to the problem of linking distributed and diverse cultural heritage information sources. In co-operation with the CIMI Consortium, Aquarelle developed and designed an architecture for an integrated resource discovery system based on the Z39.50 protocol, producing a new Z39.50 Application Profile (Michard, et al., 1998). The project also created an environment for the creation, dissemination and maintenance of documents called 'folders' that would contain secondary information about cultural heritage resources (Doerr, Fundulaki and Christophidis, 1997).

3.2 The Art Museum Image Consortium (AMICO)

The Art Museum Image Consortium (AMICO) is a not-for-profit organization of institutions that have collections of art. AMICO members are building 'The AMICO Library,' a resource that makes digitised versions of artworks available to educational institutions and other museums by subscription (Trant, Bearman & Richmond, 2000).

Each work that is included in The AMICO Library requires a description (catalog record), a digital image (adhering to a specification defined by the consortium) and a metadata record for the image itself. A 'Data Dictionary' defines the data required for both catalog record and image metadata. This is published as part of the AMICO Data Specification, currently in version 1.3 (http://www.amico.org/AMICOlibrary/dataDictionary.html).

This metadata schema can record extremely detailed information about the type of object being described, its title, physical characteristics, creation, meaning, exhibition history, ownership history (including rights) and relations. The metadata for the associated image (or other types of media) includes some fields that are based on DCMES, as well as administrative metadata.

The AMICO Data Dictionary is very detailed, so the following list only includes "core" fields that are mandatory for an AMICO record (Table 2). The Data Dictionary also includes lists of authorised terms and schemes (e.g. the Art and Architecture Thesaurus or the Library of Congress Thesaurus of Graphic Materials), basic content guidelines and some examples. Some content is provided by AMICO itself or can be generated automatically.

Table 2: AMICO Data Dictionary, v. 1.3 (core elements only)

Source: http://www.amico.org/AMICOlibrary/dataDictionary.html

More information on AMICO can be found at: http://www.amico.org/

3.3 The CIMI Consortium

Several important metadata initiatives with direct relevance to these issues have originated in the cultural heritage sector, primarily from the museum community. Investigation into standards for the interchange of cultural heritage information has been co-ordinated by the CIMI Consortium, (formerly the Consortium for the Computer Interchange of Museum Information). CIMI aims to provide demonstrations of how distributed and heterogeneous data can be accessed in a consistent way. One of CIMI's first projects was CHIO (Cultural Heritage Information Online). This project offered a demonstrator that explored the use of the Standard Generalised Markup Language (SGML) and the utility of the Z39.50 protocol (Moen, 1998). The project resulted in the publication of a CIMI Z39.50 Application Profile for cultural heritage information (CIMI Z39.50 Working Group, 1998). Other projects included a Dublin Core testbed that explored the use of DC qualifiers, the emerging DC data model and the use of RDF.

Recent CIMI (http://www.cimi.org/test_beds.html) initiatives have included the testing of the mda's SPECTRUM XML DTD, an initiative called 'Handscape' that is investigating the use of wireless technologies for museum visitors and a CIMI-based test of the OAI metadata harvesting protocol.

More information on CIMI can be found at: http://www.cimi.org/

3.4 Electronic Library Image Service for Europe (ELISE)

The Electronic Library Image Service for Europe (ELISE) was two EU-funded projects that investigated access to distributed and heterogeneous information objects, with specific reference to digital images (Eyre, 1997; 1998). The project was largely concerned with specifying a sustainable digital image service - including support for user registration and validation, rights management and charging mechanisms. A demonstration system (based on Z39.50) was built in order to show how such a service might work. Digitised content came from project partners, which included (in phase II of the project) the Victoria and Albert Museum, the Hunt Museum (Limerick) and a Dutch biomedical laboratory.

More information on ELISE can be found at: http://severn.dmu.ac.uk/elise/

3.5 Museum Educational Site Licensing Project (MESL)

The Museum Educational Site Licensing Project (MESL) was a collaboration between museums and educational institutions in the United States, with support from the Getty Art History Information Program and MUSE Educational Media, to investigate the capture, distribution and educational use of digital images and their corresponding text descriptions (Trant, 1997). The project acknowledged the uncertainty that exists with regard to intellectual property rights - particularly for image-based material - and attempted to bring stakeholders together to explore the technical, administrative and legal issues surrounding the educational use of networked digital images. As part of this, MESL developed a means of searching across diverse metadata by mapping these to a data dictionary with thirty-two fields (Blackaby and Sandore, 1997).

Table 3: MESL Data Dictionary (draft, 1996)

Source: http://lists.village.virginia.edu/uvamesl/search_help/datadict.html

3.6 The Scottish Cultural Resources Access Network (SCRAN)

The Scottish Cultural Resources Access Network (SCRAN) is a 'resource base' made up of selected multimedia objects that have been selected and digitised from the holdings of cultural heritage organisations in Scotland. Royan (1998) notes that SCRAN acts as a metadata repository, directing users "to digitised assets in its own resource base, but also to millions of undigitised objects in museums and libraries, and to digital records in other electronic collections." Development of the SCRAN Data Standard drew on both the cross-domain DCMES and SPECTRUM: the UK Museum Documentation Standard (Morrison, 1998).

The metadata SCRAN uses must not be constrained by the terminology or conventions of any one of the many domains SCRAN materials have been derived from, whether libraries, museums, archives or archaeological corpora. It must be at a significantly general level of definition to be helpful to the non-specialist user, and hospitable for searching across domains (Royan, 1998).

SCRAN resources are stored in a database system based on System Simulation's Index+ software toolkit. There are two types of search, a Quick Search with a single search box and an Assisted Search which allows more advanced searching under the headings: 'what,' 'where,' 'who,' 'when,' and 'subject' with filters for type of display, record types and sort options. Morrison (1998) said that 'what,' 'where,' 'who' and 'when' type questions had been shown to be the most frequent access points for cultural heritage information, both by early users of SCRAN and by older projects.

Record displays typically have a title, a statement of responsibility, an image thumbnail (where appropriate) and a three paragraph 'caption.' Also visible are the record details (metadata), which include some of the following elements:

Table 4: SCRAN Data Standard elements

Source: http://www.scran.ac.uk/

Subject information appears to differ, depending on the provenance of the resource. Some have subject terms taken from the Library of Congress Subject Headings (LCSH), while others appear to have classification codes or accession numbers, e.g. the photographic images from the Valentine Collection at St Andrews University Library. Fields are repeatable, and can be qualified. For example, the date can refer to both when an item was first made and when it was acquired by the collecting institution. The 'Who' element can record both artists and the provenance of an item. Events can refer to all of the various exhibitions items have been part of.

SCRAN has also implemented the Z39.50 protocol, so that the resource base can be searched in parallel with other Z39.50 'Targets.'

More information on SCRAN can be found at: http://www.scran.ac.uk/

3.7 SPECTRUM: the UK Museum Documentation Standard

SPECTRUM: the UK museum documentation standard (mda, 1997) resulted from a collaboration of over one hundred practitioners working in the area of documentation in museums. The standard was co-ordinated by the mda (formerly the Museum Documentation Association).

SPECTRUM comprises procedures for documenting objects and the processes that they undergo. It also identifies and describes the information that needs to be maintained to support those procedures. The intention is that the standard should contain all those functions that are common to most museums. A particular institution would then choose and use those procedures that are most relevant to its own requirements. The advantage of adhering to SPECTRUM is that data exchange between organisations becomes much more feasible.

The mda, in collaboration with the CIMI Consortium (see also section 3.3) and other organisations, has developed a XML Document Type Description (DTD) for the SPECTRUM standard. CIMI is currently leading a test-bed initiative among its members in order to test the DTD in practical applications. It is hoped that the existence of a standardised DTD will facilitate the interchange of collections-based information between museums (Degenhart Drenth, 2001).

This list covers all the different units of information in SPECTRUM. The units of can be used to describe the object itself as well as the different events in its history, for example its production, collection, acquisition and exhibition history.

Table 5: SPECTRUM Units of Information

Source: mda (1997).

More information on SPECTRUM can be found at: http://www.mda.org.uk/spectrum.htm

3.8 Visual Resources Association (VRA) Core Categories

The Visual Resources Association (VRA) are a group of image media professionals, including museum curators, art and architecture librarians and a wide range of other professions with an interest in the management of images (http://www.vraweb.org/)

In 2000, the VRA Data Standards Committee defined the VRA Core Categories, version 3.0. This is an element set that is intended for the creation of records to describe "works of visual culture as well as the images that represent them" (http://www.vraweb.org/vracore3.htm). The set is heavily influenced by the Dublin Core Metadata Element Set (DCMI, 1999). One of the major objectives of VRA Core 3.0 is to facilitate the sharing of information among visual resources collections about both works and their corresponding images. VRA Core 3.0 acknowledges that the elements provided might not be sufficient to fully describe a local collection and accepts that additional fields may be required to supplement the set.

Table 6: VRA Core 3.0

Source: http://www.vraweb.org/vracore3.htm

More information on VRA Core can be found at: http://www.vraweb.org/vracore3.htm

4. Digitisation initiatives

Organisations and projects concerned with digitisation have also needed to address metadata issues. It has long been recognised that collections of digitised objects - especially digital images - need metadata in order that users can access and understand them. Without metadata, digitised images would just consist of a set of files stored in a disk directory or set of directories. At the very least, digitised objects need some descriptive metadata. In addition, an object may also require information on its source or provenance, any terms and conditions associated with it (e.g. rights metadata), how the data might best be managed (e.g. administrative metadata) and how it may be linked to other objects.

A number of initiatives have begun to develop metadata standards that support the varied requirements of digitisation initiatives. One strand of research has focused on the development of standard object models that would enhance the use of digitised objects across distributed repositories. Perhaps the best example of this approach is the digital library object model developed by the Making of America II Testbed Project (MOA2) followed by the METS (Metadata Encoding & Transmission Standard) initiative. A second strand has concentrated on defining the descriptive data elements that may be considered necessary for long-term preservation. Examples of these are the report of the Research Libraries Group Working Group on the Preservation Issues of Metadata (RLG, 1998) and the NISO draft Data dictionary: technical metadata for digital still images (NISO, 2000).

The MOA2 Project (Hurley, et al., 1999, p. 7) defined three main types of metadata that can aid the discovery, navigation and management of resources:

• Descriptive metadata, used for the discovery and identification of an object.
• Structural metadata, used to display and navigate an object.
• Administrative metadata, the management information for the object.

4.1 The METS initiative

The Metadata Encoding & Transmission Standard (METS) initiative is attempting to provide an XML-based document format for encoding metadata to aid the management and exchange of digital library objects. The initiative has adapted the XML Document Type Definition developed by MOA2 and defined it in the XML schema language. The schema defined by the METS initiative separates metadata into four sections. These are 'descriptive metadata,' 'administrative metadata,' 'file groups' and 'structural maps,' the last two of which are intended to group together all of the files that make up a particular digital object and to link content and metadata to a particular structure. The administrative metadata section is intended to store technical information about the file, as well as information about intellectual property rights held in the resource, the source material, and provenance metadata that records relationships between files and migrations. Broadly speaking, the METS schema provides an XML-based container that could be used to store much of the metadata defined in preservation metadata specifications like that published by the Cedars project (Russell, et al., 2000).

More information on METS can be found at: http://www.loc.gov/standards/mets/

4. The NISO draft Technical Metadata for Digital Still Images

The Data dictionary: technical metadata for digital still images is a draft NISO (National Information Standards Organization) standard currently under review (NISO, 2000). Development of the draft standard grew out of an 'Image Metadata Workshop' held in 1999, sponsored by NISO, the Council for Library and Information Resources (CLIR) and the Research Libraries Group (RLG). The draft standard is not intended to duplicate work on descriptive metadata schemas, but to help define a standardised way of recording the technical attributes of digital images and the production techniques associated with them. The data dictionary includes elements that will record detailed information about images themselves (e.g. formats, compression, etc.), the image creation process, some quality metrics, and any change history (e.g. migrations). The draft is mainly concerned with semantics and no particular syntax is proposed or recommended. Development of the draft standard is based on the experiences of digitisation centres. If and when it is adopted as a standard, it will be of particular use for helping to support the long-term preservation of the products of digital imaging projects.

Elements defined in the draft data dictionary include those that record basic information about formats (e.g. 'MIMEType,' 'Format,' 'Compression') and files (e.g. 'FileSize,' 'Checksum'), the image creation process (e.g. 'SourceType,' 'ScanningAgency,' 'DateTimeCreated,' etc.), image performance (associated with image quality) and change history.

More information on the NISO Technical Metadata draft can be found at: http://www.niso.org/committees/committee_au.html

5. Geospatial metadata initiatives

Geospatial image types - e.g. maps, geographical information systems, etc. - have specific resource description requirements that have been addressed by the development of separate standards. The most important of these are initiatives of the US Federal Geographic Data Committee (FGDC) and the International Organization for Standardization (ISO).

5.1 FGDC Content Standard for Digital Geospatial Metadata

The Content Standard for Digital Geospatial Metadata (CSDGM) was developed by the FGDC in order to provide a common set of terminology and definitions for the documentation of digital geospatial data. Work on developing CSDGM commenced in 1992 and Version 1 of the standard was approved by the FGDC in June 1994. Version 2 was approved in June 1998 and provides more flexibility with regard to user profiling and extensibility (FGDC, 1998).

The FGDC standard can be used to describe all types of geospatial data, not just images. It is a fairly large schema, with over 200 elements, but a number of metadata creation tools have been developed by various agencies (http://www.fgdc.gov/clearinghouse/mitre/task2/tools.html).

More information on FGDC CSDGM can be found at: http://www.fgdc.gov/metadata/csdgm/

5.2 ISO 19115 Geographical information -- Metadata

Related to the CSGDM is the standard for geographic information that is being developed by ISO's Technical Committee 211 (ISO/TC 211). This committee is developing a series of standards (the ISO 19100 series) to help facilitate the description and management of geographic information (Østensen, 2001). ISO/DIS 19115 provides a metadata schema for describing digital geographical datasets. This became an ISO Draft International Standard in mid 2001 and has already begun to be implemented, e.g. by ANZLIC - the spatial information council of Australia and New Zealand (http://www.anzlic.org.au/asdi/metagrp.htm) and in tools like ArcCatalog (Vienneau & Danko, 2001).

There is a hope that most existing (and future) standards development for geospatial metadata will be able to converge through ISO 19115. Versions of the ISO 19115 schema have already been mapped to the ANZLIC Metadata Guidelines (ANZLIC, 2001) and the Directory Interchange Format (DIF) used by NASA's Global Change Master Directory (Solomon, 2000). The FGDC are also involved in the development of ISO 19115 and are committed to harmonising the CSGDM with the ISO standard (http://www.fgdc.gov/metadata/whatsnew/fgdciso.html).

More information on ISO 19115 can be found at: http://www.isotc211.org/

6. Educational metadata initiatives

The growing use of information technology to facilitate learning processes (e.g. lifelong learning) has led to an interest in descriptive data about learning resources. Initiatives like the Gateway to Educational Material (GEM), the Education Network of Australia (EdNA) and the European Schoolnet (http://www.eun.org/) have set up services that give access to educational materials. There is, therefore, a good deal of interest in the development of standardised metadata that can describe learning resources. An early attempt to develop an educational metadata specification was made by the EU-funded ARIADNE projects (http://www.ariadne-eu.org/). Both ARIADNE and the IMS consortium (see 6.2 below) have fed into the development of the IEEE Learning Object Metadata (LOM) standard, and both initiatives' current metadata specifications are derived from it.

There is a wide range of educational metadata standards in use. Apart from LOM, initiatives like EdNA and European Schoolnet have implemented metadata based on the DCMES with various refinements and extensions. An UK initiative, known as the Metadata for Education Group (MEG), is attempting to reach consensus on educational metadata, and the MEG Concord (http://www.ukoln.ac.uk/metadata/education/documents/concord.html) currently has over 70 signatories.

More information on MEG can be found at: http://www.ukoln.ac.uk/metadata/education/

6.1 Draft Standard for Learning Object Metadata (LOM)

The Draft Standard for Learning Object Metadata is an initiative of the IEEE (Institution of Electrical and Electronics Engineers) Learning Technology Standards Committee (LTSC). The standard is currently in draft 6.3 (18 January 2002), but it is hoped that it will shortly become a full IEEE standard (Duval, 2001). Its development builds on the experiences of the ARIADNE projects and the IMS consortium, and has close links with DCMI. A joint 'Memorandum of Understanding' was agreed between the LTSC LOM working group and DCMI in December 2000 to establish a co-ordinated discussion between the two initiatives (http://dublincore.org/news/pr-20001206.shtml). The LTSC draft also includes a mapping from the 15 DC elements to the LOM draft standard.

The LOM standard is a fairly large element set (IEEE LTSC, 2002) grouped into several categories, e.g. general, lifecycle, administrative (here called 'meta-metadata'), technical, educational, rights, etc. The general category elements are used to describe a learning object as a whole and many elements are based on DCMES, but with additional elements for structure and aggregation level - acknowledging that learning objects can be found at several different levels of granularity. For example, most of the images in the proposed FILTER database would have a LOM aggregation level of '1,' while the exemplar learning resources would normally be at levels '2' or '3.' The technical category contains elements for Format (e.g., Internet MIME types), size, location, technical requirements for use, installation remarks, hardware and software requirements, etc. Under the educational category, there are elements for the type and level of interactivity in a learning object, the type of learning resource, the intended user, age range, learning time, etc. Relevant terms are normally taken from a fixed vocabulary defined as part of the LOM standard, e.g. examples of learning resource types might be: questionnaire, narrative text, exam or experiment, etc. The relation category enables links to be made between learning objects, e.g. in FILTER between the exemplar learning resources and individual images. Subject notation and terms can be recorded as part of the classification category, as can a number of other taxonomies.

More information on LOM can be found at: http://ltsc.ieee.org/wg12/

6.2 The IMS Meta-data Specification

IMS Global Learning Consortium, Inc. (originally the Instructional Management Systems project) is a membership organisation that develops and promotes open specifications for facilitating online learning. These include an IMS Content Packaging Specification that enables the combination of learning objects into interoperable packages and an IMS Meta-data Specification. Version 1 of the metadata specification was published in 1999, and the current version (1.2) in June 2001. Version 1.2 of the IMS specification follows draft 6.1 of the LTSC LOM standard but has made a few minor changes of element names, and is extensible. The full metadata specification includes an Information Model that defines the LOM elements (and IMS modifications), an XML Binding specification that proposes a syntax, and a Best Practice and Implementation Guide to provide general guidance about how an application may use LOM metadata elements.

More information on IMS can be found at: http://www.imsproject.org/

6.3 DCMI Education Working Group

The DCMES has been used by a number of services that describe educational resources, including GEM and EdNA. Their experiences suggested that it would be a good idea to set up a working group of the DCMI to discuss and develop a proposal for the use of DC metadata in the description of educational resources. As a result, the Education Working Group of the DCMI was formed in 1999. In 2000, this working group produced a DCMI Draft Proposal (Mason & Sutton, 2000) that included new elements for 'Audience' and 'Standard' together with endorsements of three data elements taken from the IEEE LOM namespace: 'InteractivityType,' 'InteractivityLevel' and 'TypicalLearningTime.' In May 2002, the DCMI Usage Board (2001) accepted 'Audience' as a domain specific element as part of the DC-Terms namespace (http://purl.org/dc/terms/). More recently, the working group have published a draft proposal for an 'audienceLevel' qualifier for the 'Audience' element (Sutton, 2002).

Application profiles of DCMES and DC-Ed namespaces have also been used in the EU-funded EASEL (Educator Access to Services in the Electronic Landscape) project (Sandford, Slavic & Cox, 2001; Slavic & Baiget, 2001).

The Web pages of the DCMI Education Working Group can be found at: http://dublincore.org/groups/education/

7. Subject schemes

There are a few subject schemes that have been specifically developed for image material, e.g. the Art & Architecture Thesaurus, the Thesaurus for Graphical Materials and ICONCLASS. These have mainly been developed to support the visual arts.

7.1 The Art & Architecture Thesaurus (AAT)

The Art & Architecture Thesaurus (AAT) was developed by the J. Paul Getty Trust and is maintained by the Getty Vocabulary Program (part of the Getty Research Institute). The version of AAT available on the Web is not directly linked to a database of the Getty's collections. On the Web interface to AAT, it is possible to browse a list of 'AAT Hierarchies' that organises descriptors (preferred terms) by one of seven high-level 'facets'. Alternatively, terms and their associated scope notes can be searched directly through the Web interface. The information available on each descriptor includes, where appropriate, a scope note, together with some additional information that might include broader or narrower terms, related terms, alternative forms of speech, British English equivalents, historical notes, etc

More information on AAT can be found at: http://www.getty.edu/research/tools/vocabulary/aat/

7.2 Thesaurus for Graphic Materials

The Thesaurus for Graphic Materials was developed to support the cataloguing and retrieval needs of the Prints and Photographs Division of the Library of Congress. The standard has two parts. Part I (TGM I) covers Subject Terms provides a list of terms that can be used for the subject indexing of pictorial materials. Part 2 (TGM II) provides a list of descriptors for the form and genre of such materials. The current edition of TGM I contains over 6,300 entries and has a close association with the Library of Congress Subject Headings (LCSH). The latest versions of TGM I exclude proper names (e.g. for people, organisations, events, structures, or geographic places). It is envisaged that users would normally want to create local or shared authority files. For example, the Prints and Photographs Division uses TGM I in conjunction with the LC Name Authority File and LCSH.

More information on TGN I can be found at: http://lcweb.loc.gov/rr/print/tgm1/

7.3 ICONCLASS

ICONCLASS is a "classification system for iconographic research and the documentation of images" (http://www.iconclass.nl/texts/icsys.htm). It was first devised by Henri van de Waal (1910-1972), Professor of Art History at the University of Leiden, and published in 17 volumes by the Koninklijke Nederlandse Akademie van Wetenschappen (Royal Netherlands Academy of Arts and Sciences) between 1973 and 1985 (Grund, 1993). In 1990-1991, the Department of Computers & Arts of Utrecht University prepared a computerised version of the ICONCLASS System: the ICONCLASS Browser. This lets users search the classification system using keywords and browse all notations. The current version, the Iconclass2000 Browser, is distributed on CD-ROM although Web-based versions of the browser are also available.

ICONCLASS is an alphanumerical classification of the subjects of Western art, offering definitions and keywords in English. With this scheme it is possible to describe objects, persons, events, situations and abstractions that appear in visual images. From 1996, it has been divided into 10 main divisions: 0. Abstract, Non-representational Art, 1. Religion and Magic, 2. Nature, 3. Human being, Man in general, 4. Society, Civilization, Culture, 5. Abstract Ideas and Concepts, 6. History, 7. Bible, 8. Literature, 9. Classical Mythology and Ancient History.

More information on ICONCLASS can be found at: http://www.iconclass.nl/

8. Conclusions

This report has attempted to describe a wide range of metadata initiatives relating to the description of images and learning resources.

With regard to image resources, it is clear that metadata is needed to support a variety of different functions. As well as some kind of discovery metadata, images also need data to be stored about technical characteristics (e.g. formats, capture processes, etc.), the original objects from which they are derived, and any IPR that may be vested in them. Naturally, the metadata initiatives described in this report have been designed for different purposes, and they differ as to the precise amount of attention they give to each of these metadata types. The data dictionary developed for use in The AMICO Library, for example, is rich in metadata about original objects, e.g. their dimensions, ownership, IPR, etc. The draft NISO Data dictionary: technical metadata for digital still images is focused more, as it's name suggests, on the technical aspects of digital images, e.g. formats, capture devices, change history, etc. Other initiatives are much simpler, and have been influenced by the development of generic metadata element sets like the DCMES. Examples of these are the VRA Core Categories and the SCRAN Data Standard. These are probably closer to fulfilling the requirements of the relatively small FILTER exemplar database.

As for learning resource metadata, much of the current standardisation effort is focused on the IEEE LTSC Draft Standard for Learning Object Metadata (IEEE LTSC, 2002). Much development work exists elsewhere, co-ordinated in the UK by the Metadata for Education Group. As with images, there is some focus on the Dublin Core. DCMES v. 1.1 with some additions from the DC-Terms namespace (e.g., DCMI Usage Board, 2001) and the Draft Standard for Learning Object Metadata (IEEE LTSC, 2002) might provide a good starting place for describing the learning resources that will be included in the FILTER exemplar database.

The image-specific subject schemes described in this report mostly relate to art and architecture (AAT, ICONCLASS) or to photographic materials (TGM I, APT). It is unlikely that any one scheme would be suitable for all of the LTSN centres involved in FILTER, which have a much wider focus on the sciences and social sciences.

Notes

1. Panofsky prefers the term 'iconographical interpretation' in his 1962 edition of Studies in Iconology (Panofsky, 1962, p. 14).
2. For a review of XML standards and tools relevant to the cultural heritage domain, see: F. Hernández, P. Ribes, G. Koch, B. Mulrenin & R. Yeates, Review of the state of the art. COVAX deliverable, D1.2, 31 March 2001: http://www.covax.org/public_docum/p_documets.htm

9. References

Andres, P.M., Amy, L., Steinwand, D., Lawson, J. & Moffitt, R.A. (1999). Making sense of space-based earth image data: the KidSat data system. IEEE Transactions on Geoscience and Remote Sensing, 37 (4), 1785-1800.

ANZLIC. (2001). A mapping between ANZLIC Metadata Guidelines and ISO/DIS 19115 Metadata Standard. Canberra: ANZLIC, 9 March.

Armitage, L.H. & Enser, P.G.B. (1997). Analysis of user need in image archives. Journal of Information Science, 23 (4), 287-299.

Arms, C.R. (1999). Getting the picture: observations from the Library of Congress on providing online access to pictorial images. Library Trends, 48 (2), 379-409.

Besser, H. (1997). Image databases: the first decade, the present, and the future. In: Heidorn, P.B. & Sandore, B., eds. Digital image access and retrieval. Urbana-Champaign, Ill.: University of Illinois at Urbana-Champaign, Graduate School of Library and Information Science, 11-28.

Besser. H. & Lack, R. (1999). Image and metadata distribution at seven university campuses: reports from a study of the Museum Educational Site Licensing Project. In: Abiteboul, S. & Vercoustre, A.-M., eds. Research and advanced technology for digital libraries: Third European Conference, ECDL '99, Paris France, September 22024, 1999, proceedings. Lecture Notes in Computer Science, 1696. Berlin: Springer.

Besser, H. & Trant, J. (1994). Describing image files: the need for a technical standard. Coalition for Networked Information, Fall Meeting, Orlando, Florida, 30 November. http://www.archimuse.com/papers/cni.fall.94.html (visited 8 March 2002).

Besser, H. & Trant, J. (1995). Introduction to imaging: issues in constructing an image database. Santa Monica, Calif.: Getty Art History Information Program.

Bewley, R., Donoghue, D., Gaffney, V., Leusen, M. van & Wise, A. (1999). Archiving aerial photography and remote sensing data: a guide to good practice. AHDS Guides to Good Practice. Oxford: Oxbow. http://ads.ahds.ac.uk/project/goodguides/apandrs/ (visited 8 March 2002).

Birdsey, C.L., Golding, A. & Jacobson, R.E. (1999). The effect of digital indexing systems on the accessibility of photographic collections. Archives and Museum Informatics, 13, 181-190.

Blackaby, J. & Sandore, B. (1997). Building integrated museum information retrieval systems: practical approaches to data organisation and access. Archives and Museum Informatics; 11, 117-146.

Bollacker, K.D., Lawrence, S. & Giles, C.L. (2000). Discovering relevant scientific literature on the Web. IEEE Intelligent Systems, 15 (2), 42-47.

Bormans, J. & Hill, K. (2001). MPEG-21 overview v.3. ISO/IEC JTC1/SC29/WG11, December. http://mpeg.telecomitalialab.com/standards/mpeg-21/mpeg-21.htm (visited 8 March 2002).

Carazo, J.M. & Stelzer, E.H.K. (1999). The BioImage Database Project: organising multidimensional biological images in an object relational database. Journal of Structural Biology, 125, 97-102.

Cawkell, A.E. (1992). Selected aspects of image-processing and management: review and future-prospects. Journal of Information Science, 18 (3), 179-192

Cawkell, A.E. (1993). Indexing collections of electronic images: a review. BLR&DD Research Review, 15. London: British Library Research & Development Department.

Chang, S.F., Sikora T. & Puri, A. (2001). Overview of the MPEG-7 Standard. IEEE Transactions on Circuits and Systems for Video Technology, 11 (6), 688-695.

Chang, S.F., Smith, J.R., Beigi, M. & Benitez, A. (1997). Visual information retrieval from large distributed online repositories. Communications of the ACM, 40 (12), 63-71.

Chang, S.K. & Hsu, A. (1992). Image information systems: where do we go from here? IEEE Transactions on Knowledge and Data Engineering, 4 (5), 431-442.

Chen, H.L. (2001). An analysis of image retrieval tasks in the field of art history. Information Processing and Management, 37, 701-720.

Chen, H.L. & Rasmussen, E.M. (1999). Intellectual access to images. Library Trends, 48 (2), 291-302.

Chen. Z., Wenyin, L., Zhang, F., Li, M. & Zhang, H. (2001). Web mining for Web image retrieval. Journal of the American Society for Information Science and Technology, 52 (10), 831-839.

Chu H.T. (2001). Research in image indexing and retrieval as reflected in the literature. Journal of the American Society for Information Science and Technology, 52 (12), 1011-1018.

CIMI Z39.50 Working Group. (1998). The CIMI Profile: Z39.50 Application Profile for Cultural Heritage Information, release 1.0H. Denton, Tex.: University of North Texas, School of Library and Information Sciences. http://www.cimi.org/public_docs/HarmonizedProfile/HarmonProfile1.htm (visited 8 March 2002).

College Art Association & Getty Research Institute. (2000). Categories for the Description of Works of Art, v. 2.0. Los Angeles, Calif.: Getty Research Institute. http://www.getty.edu/research/institute/standards/cdwa/ (visited 8 March 2002).

Constantopoulos, P. & Doerr, M. (1995). An approach to indexing annotated images. In: Bearman, D., ed. Multimedia computing and museums: selected papers from the Third International Conference on Hypermedia and Interactivity in Museums (ICHIM '95 - MCN '95), San Diego, California, October 9-13, 1995. Pittsburgh, Pa.: Archives & Museum Informatics, 278-298. Also available: http://citeseer.nj.nec.com/constantopoulos95approach.html (visited 8 March 2002).

Couprie, L.D. (1983). ICONCLASS. Art Libraries Journal, 8 (2), 32-49.

Degenhart Drenth, B. (2001). Building on the mda SPECTRUM-XML DTD for collections management data interchange. In: Museums and the Web 2001 [proceedings]. Pittsburgh, Pa.: Archives & Museum Informatics. http://www.archimuse.com/mw2001/papers/degenhart/degenhart.html (visited 8 March 2002).

Dempsey, L. & Heery, R. (1998). Metadata: a current view of practice and issues. Journal of Documentation, 54 (2), 145-172.

Dodson, H., Levin, P., Ride, S. & Souviney, R. (2000). The EarthKAM project: creating space imaging tools for teaching and learning. Computers & Geosciences, 26, 683-691.

Doerr, M., Fundulaki, I. & Christophidis V. (1997). The specialist seeks expert views: managing digital folders in the AQUARELLE project. In: Museums and the Web 97 [proceedings]. Pittsburgh, Pa.: Archives & Museum Informatics. http://www.archimuse.com/mw97/speak/doerr.htm (visited 8 March 2002).

Dublin Core Metadata Initiative. (1999). Dublin Core Metadata Element Set, Version 1.1: Reference Description. DCMI Recommendation, 2 July. http://dublincore.org/documents/dces/ (visited 8 March 2002).

Dublin Core Metadata Initiative. (2000). Dublin Core qualifiers. DCMI Recommendation, 11 July. http://dublincore.org/documents/dcmes-qualifiers/ (visited 8 March 2002).

DCMI Usage Board. (2001). DC Usage Board Decision 1 (DC-Education Proposal). 21 May. http://dublincore.org/usage/decisions/2001/education-01.shtml (visited 8 March 2002).

Duffy, C. (1997). An introduction to the Performing Arts Data Service. Literary and Linguistic Computing, 12 (4), 277-281.

Duval, E. (2001). Appendix A: educational domain. In: Eilert, C. & Dekkers, M., Metadata Watch report #6 and Standards Framework report #3. SCHEMAS project, 11 November. http://www.schemas-forum.org/metadata-watch/sixth/sectiona.html (visited 8 March 2002).

Eakins, J.P. (1998). Techniques for image retrieval. Library & Information Briefings, 85. London: South Bank University, Library Information Technology Centre.

Eakins, J.P. (2002). Towards intelligent image retrieval. Pattern Recognition, 35, 3-14.

Eakins, J.P., Boardman, J.M. & Graham, M.E. (1998). Similarity retrieval of trademark images. IEEE Multimedia, 5 (2), 53-63.

Eakins, J. & Graham, M. (1999). Content-based image retrieval. JTAP Report 39. Manchester: JISC Technology Applications Programme. http://www.unn.ac.uk/iidr/research/cbir/report.html (visited 8 March 2002).

Enser, P.G.B. (1995). Progress in documentation: pictorial information retrieval. Journal of Documentation, 51, 126-170.

Enser, P. (2000). Visual image retrieval: seeking the alliance of concept-based and content-based paradigms. Journal of Information Science, 26 (4), 199-210.

Enser, P.G.B. & Orr, H.P. (1993). Creative picture research using the Hulton Deutsch CD collection. Program, 27 (4), 389-398.

Evans, H. (1992). Practical picture research: a guide to current practice, procedure, techniques and resources. London: Blueprint.

Eyre, J. (1997). Architecture for online museums of the future: a object server for the future (ELISE II). In: Museums and the Web 97 [proceedings]. Pittsburgh, Pa.: Archives & Museum Informatics. http://www.archimuse.com/mw97/speak/eyre.htm (visited 8 March 2002).

Eyre, J. (1998). Distributed image services. VINE, 107, 65-72.

Federal Geographic Data Committee. (1998). Content standard for digital geospatial metadata. FGDC-STD-001-1998. Washington, D.C.: Federal Geographic Data Committee, June 1998. http://www.fgdc.gov/metadata/csdgm/ (visited 8 March 2002).

Ferguson, E.S. (1992). Engineering and the mind's eye. Cambridge, Mass.: MIT Press.

Flickner, M., Sawhney, H., Niblack, W., Ashley, J., Huang, Q., Dom, B., Gorkani, M., Hafner, J., Lee, D., Petkovic, D., Steele, D. & Yanker, P. (1995). Query by Image and Video Content: the QBIC system. Computer, 28 (9), September, 23-32.

Gill, T. & Grout, C. (1997). Finding and preserving visual arts information on the Internet. Art Libraries Journal, 22 (3), 19-25.

Gillings, M. & Wise, A. (1999). GIS guide to good practice. AHDS Guides to Good Practice. Oxford: Oxbow. http://ads.ahds.ac.uk/project/goodguides/gis/ (visited 8 March 2002).

Gonzalez-Couto, E., Hayes, B. & Danckaert, A. (2001). The life sciences Global Image Database (GID). Nucleic Acids Research, 29 (1), 336-339.

Goodrum, A.A. (2000). Image information retrieval: an overview of current research. Informing Science, 3 (2). http://inform.nu/Articles/Vol3/indexv3n2.htm (visited 8 March 2002).

Goodrum, A. & Spink, A. (2001). Image searching on the Excite Web search engine. Information Processing and Management, 37, 295-311.

Graham, P.S. (1994). Intellectual preservation: electronic preservation of the third kind. Washington, D.C.: Commission on Preservation and Access. http://www.clir.org/pubs/abstract/pub45.html (visited 8 March 2002).

Graham, M. & Eakins, J. (1998). ARTISAN: a prototype retrieval system for trade mark images. VINE, 107, 73-80.

Grosky, W.I. (1997). Managing multimedia information in database systems. Communications of the ACM, 40 (12), 73-80.

Grout, C., Purdy, P., Youngs, K., Williams, J., Lock, A., Brickley, D. & Moss, O. (2000). Creating digital resources for the visual arts: standards and good practice. AHDS Guides to Good Practice. Oxford: Oxbow. http://vads.ahds.ac.uk/guides/creating_guide/contents.html (visited 8 March 2002).

Grund, A. (1993). ICONCLASS: on subject analysis of iconographic representations of works of art. Knowledge Organisation, 20 (1), 20-29.

Gudivada, V.N. & Raghavan, V.V. (1995). Content based image retrieval systems. Computer, 28 (9), 18-22.

Gudivada, V.N. & Raghavan, V.V. (1997). Modeling and retrieving images by content. Information Processing & Management, 33 (4), 427-452.

Hunter, J. (1999). MPEG-7: behind the scenes. D-Lib Magazine, 5 (9), September. http://www.dlib.org/dlib/september99/hunter/09hunter.html (visited 8 March 2002).

Hunter, J. (2001). An overview of the MPEG-7 Description Definition Language (DDL). IEEE Transactions on Circuits and Systems for Video Technology, 11 (6), 765-772.

Hunter, J. & Armstrong, L. (1999). A comparison of schemas for video metadata representation. Computer Networks, 31, 1431-1451.

Hunter, J. & Nack, F. (2000). An overview of the MPEG-7 Description Definition Language (DDL) proposals. Signal Processing: Image Communication, 16, 271-293.

Hurley, B.J., Price-Wilkin, J., Proffitt, M. & Besser, H. (1999). The Making of America II Testbed Project: a digital library service model. Washington, D.C.: Council on Library and Information Resources. http://www.clir.org/pubs/abstract/pub87abst.html (visited 8 March 2002).

IEEE Learning Technology Standards Committee (2002). Draft standard for learning object metadata, draft 6.3, 18 January. Latest draft available from: http://ltsc.ieee.org/wg12/ (visited 8 March 2002).

ISO/DIS 19115 (2001). Geographic information - Metadata. ISO Technical Committee 211. http://www.isotc211.org/ (visited 8 March 2002).

Jörgensen, C. (1998). Attributes of images in describing tasks. Information Processing & Management, 34 (2/3), 161-174.

Jörgensen, C. (1999) Access to pictorial material: a review of current research and future prospects. Computers and the Humanities, 33 (4), 293-318.

Kenney, A.R. & Rieger, O.Y. (2000). Moving theory into practice: digital imaging for libraries and archives. Mountain View, Calif.: Research Libraries Group.

Kobayashi, M. & Takeda, K. (2000). Information retrieval on the Web. ACM Computing Surveys, 32 (2), 144-173.

Koenen, R. (2001). Object-based MPEG offers flexibility. EE Times, 12 November. http://www.eetimes.com/story/OEG20011112S0042 (visited 8 March 2002)

Koenen, R. & Pereira, F. (2000). MPEG-7: a standardised description of audiovisual content. Signal Processing: Image Communication, 16, 5-13.

Lawrence, S. & Giles, C.L. (1998a). Searching the World Wide Web. Science, 280, 98-100.

Lawrence, S. & Giles, C.L. (1998b). Context and page analysis for improved Web search. IEEE Internet Computing, 2 (4), 38-46.

Lawrence, S. & Giles, C.L. (1999a). Searching the Web: general and scientific information access. IEEE Communications Magazine, 37 (1), 116-122.

Lawrence, S. & Giles, C.L. (1999b). Accessibility of information on the Web. Nature, 400, 107-109.

Lynch, C. (1997). Searching the Internet. Scientific American, 276 (3), 44-48. http://www.sciam.com/0397issue/0397lynch.html (visited 8 March 2002).

Maier, E. (1999). Will online images prove to be a multimedia cornucopia for museums? Aslib Proceedings, 51 (10), 325-334.

Mason, J. & Sutton, S. (2000). Education Working Group: draft proposal. DCMI Working Draft. http://dublincore.org/documents/education-namespace/ (visited 8 March 2002).

McLean, S., Spring, M.B., Rasmussen, E. & Williams, J.G. (1995). Online image databases: usability and performance. The Electronic Library, 13 (1), 27-41.

Museum Educational Site Licensing Project. (1996). MESL Data Dictionary, v. 1.0, 2nd draft, 23 January. http://lists.village.virginia.edu/uvamesl/search_help/datadict.html (visited 8 March 2002).

Michard, A., Christophides, V., Scholl, M., Stapleton, M., Sutcliffe, D., Vercoustre, A.-M. (1998). The Aquarelle resource discovery system. Computer Networks and ISDN Systems, 30 (13), 1185-1200.

Moen, W.E. (1998). Accessing distributed cultural heritage information. Communications of the ACM, 41 (4), 45-48.

Morrison, I. (1998). SPECTRUM and the SCRAN data standard. mda Information, 3 (1). http://www.mda.org.uk/info31im.htm (visited 8 March 2002).

Museum Documentation Association. (1997). SPECTRUM: the UK museum documentation standard, 2nd ed. Cambridge: Museum Documentation Association.

Nack, F. & Lindsay, A.T. (1999a). Everything you wanted to know about MPEG-7: part 1. IEEE Multimedia, 6 (3), 65-77.

Nack, F. & Lindsay, A.T. (1999b). Everything you wanted to know about MPEG-7: part 2. IEEE Multimedia, 6 (4), 64-73.

National Information Standards Organization. (2000). Data dictionary: technical metadata for digital still images. Bethesda, Md.: NISO. http://www.niso.org/committees/committee_au.html (visited 8 March 2002).

Nentwig, L. (1999). LeMO: a virtual exhibition of 20th century German history. Computer Networks, 31, 2237-2244.

Østensen, O. (2001). The expanding agenda of geographic information standards. ISO Bulletin, July. http://www.iso.ch/iso/en/commcentre/isobulletin/articles/2001/Articlesindex.html (visited 8 March 2002).

Panofsky, E. (1955). Meaning in the visual arts: papers in and on art history. Garden City, N.Y.: Doubleday Anchor.

Panofsky, E. (1962). Studies in iconology: humanistic themes in the art of the Renaissance. New York: Harper & Row.

Petersen, T. (1994). Art & Architecture Thesaurus, 2nd ed. New York: Oxford University Press, on behalf of the Getty Art History Information Program.

Rasmussen, E.M. (1997). Indexing images. Annual Review of Information Science and Technology, 32, 169-196.

Research Libraries Group. (1998). RLG Working Group on Preservation Issues of Metadata: final report. Mountain View, Calif.: Research Libraries Group. http://www.rlg.org/preserv/presmeta.html (visited 8 March 2002).

Rowe, N.C. (1999). Precise and efficient retrieval of captioned images: the MARIE project. Library Trends, 48 (2), 475-495.

Royan, B. (1998). Cross-domain access to digitised cultural resources: the SCRAN project, 64th IFLA General Conference, Amsterdam, 16-21 August 1998. http://ifla.inist.fr/IV/ifla64/039-109e.htm (visited 8 March 2002).

Rui, Y., Huang, T.S. & Chang, S.-F. (1999). Image retrieval: current techniques, promising directions, and open issues. Journal of Visual Communication and Image Representation, 10, 39-62.

Russell, K., Sergeant, D., Stone, A., Weinberger, E. & Day, M. (2000). Metadata for digital preservation: the Cedars project outline specification, Leeds: Cedars project. http://www.leeds.ac.uk/cedars/metadata.html (visited 8 March 2002).

Sandford, P., Slavic, A. & Cox, A. (2001). Access to distributed learning material: the EASEL solution. VINE, 122, 65-70.

Shatford, S. (1986). Analysing the subject of a picture: a theoretical approach. Cataloguing & Classification Quarterly, 5 (3), 39-61.

Shatford Layne, S. (1994). Some issues in the indexing of images. Journal of the American Society for Information Science, 45 (8), 583-588.

Sikora, T. (2001). The MPEG-7 Visual Standard for content description - an overview. IEEE Transactions on Circuits and Systems for Video Technology, 11 (6), 696-702.

Slavic, A. & Baiget, C. (2001). Using Dublin Core in educational material: some practical considerations based on the EASEL experience. VINE, 125, 74-85.

Smith, J.R. & Chang, S.F. (1997). Visually searching the Web for content. IEEE Multimedia, 4 (3), July-September, 12-20.

Solomon, C.C. (2000). DIF to ISO Mapping. Greenbelt, Md.: NASA Goddard Space Flight Center. http://gcmd.gsfc.nasa.gov/Aboutus/standards/difiso.html (visited 8 March 2002).

Stone, H.S. (1999). Image libraries and the Internet. IEEE Communications Magazine, 37 (1), 99-106.

Sutton, S.A. (1998). Gateway to Educational Materials (GEM): metadata for networked information discovery and retrieval. Computer Networks and ISDN Systems, 30 (1-7), 691-693.

Sutton, S. (2002). DCMI Education Working Group: proposal for audienceLevel qualifier for the Audience element. http://dublincore.org/groups/education/Audience-Level-Proposal.shtml (visited 8 March 2002).

Svenonius, E. (1994). Access to nonbook materials: the limits of subject indexing for visual and aural languages. Journal of the American Society for Information Science, 45 (8), 600-606.

Swain, M.J. & Ballard, D.H. (1991). Color indexing. International Journal of Computer Vision, 7 (1), 11-32.

Townsend, S., Chappell, C. & Struijvé, O. (1999). Digitising history. AHDS Guides to Good Practice. Oxford: Oxbow. http://hds.essex.ac.uk/g2gp/digitising_history/index.asp (visited 8 March 2002).

Trant, J. (1997). Exploring new models for administering intellectual property: the Museum Educational Site Licensing Project. In: Heidorn, P.B. & Sandore, B., eds., Digital image access and retrieval. Urbana-Champaign, Ill.: University of Illinois at Urbana-Champaign, Graduate School of Library and Information Science, 29-41.

Trant, J., Bearman, D. & Richmond, K. (2000). Collaborative cultural resource creation: the example of the Art Museum Image Consortium. In: Bearman, D. & Trant, J., eds., Museums and the Web 2000: selected papers from an international conference. Pittsburgh, Pa.: Archives & Museum Informatics, 39-50. http://www.archimuse.com/mw2000/papers/trant/trant.html(visited 8 March 2001).

Vercoustre, A.-M. & Paradis, F. (1999). Metadata for photographs: from digital library to multimedia application. In: Abiteboul, S. & Vercoustre, A.-M., eds. Research and advanced technology for digital libraries: third European conference, ECDL'99, Paris, France, September 22-24, 1999, proceedings. Lecture Notes in Computer Science, 1696. Berlin: Springer, 39-57.

Vienneau, A. & Danko, D. (2001). Implementing ISO 19115. ISO/TC 211 Workshop on standards in action, Adelaide, Australia, 24 October. http://www.isotc211.org/WorkshopAdelaide/AdelaideWorkshop.htm (visited 8 March 2002).

Visual Resources Association Data Standards Committee. (2000). VRA Core Categories, version 3.0. http://www.vraweb.org/vracore3.htm (visited 8 March 2002).

Wactlar, H.D., Kanade, T., Smith, M.A. & Stevens, S.M. (1996). Intelligent access to digital video: Informedia project. Computer, May 1996.

Weibel, S. (1999). The state of the Dublin Core, April 1999. D-Lib Magazine, 5 (4), April. http://www.dlib.org/dlib/april99/04weibel.html (visited 8 March 2002).

Weibel, S. & Koch, T. (2000). The Dublin Core Metadata Initiative: mission, current activities, and future directions. D-Lib Magazine, 6 (12), December. http://www.dlib.org/dlib/december00/weibel/12weibel.html (visited 8 March 2002).

Weibel, S.L. & Lagoze, C. (1997). An element set to support resource discovery: the state of the Dublin Core, January 1997. International Journal on Digital Libraries, 1, 176-186.

Weibel, S. & Miller, E. (1997). Image description on the Internet: a summary of the CNI/OCLC Image Metadata Workshop. D-Lib Magazine, January. http://www.dlib.org/dlib/january97/oclc/01weibel.html (visited 8 March 2002).

Weibel, S., Godby, J., Miller, E. & Daniel, R. (1995). OCLC/NCSA Metadata Workshop report. Dublin, Ohio: Online Computer Library Center. http://dublincore.org/workshops/dc1/report.shtml (visited 8 March 2002).

Wong, S.T.C. & Tjandra, D.A. (1999). A digital library for biomedical imaging on the Internet. IEEE Communications Magazine, 37 (1), 84-91.

The FILTER project The FILTER project seeks to educate the tertiary education community in the use of digital images and related metadata for learning and teaching purposes. The project is funded as part of the DNER Development Programme and is co-ordinated by the Institute for Learning and Research Technology (ILRT) at the University of Bristol. More detailed information about FILTER is available from the project's Web pages at: http://www.filter.ac.uk/

Maintained by: Michael Day, Research Officer, UKOLN, University of Bath. Created: 08-Mar-2002 Last updated: 21-Mar-2002.