Report on UCISA-SG Conference

Introduction

Computer graphics, image processing and visualization software and techniques can assist in the understanding of information as well as in its presentation. The conference looked at examples of the use of such software and techniques taken from a wide range of disciplines. Papers also looked into the future at the potential of virtual reality software to assist in extending our knowledge. These examples are relevant in both teaching and research.

The conference also looked at examples of the way graphics is used in three ways:

As a gateway to information
We can examine data through a graphical gateway, this is a particularly useful technique for looking at sensitive data where the underlying data is not available to us as in the census data. The graphical interface can enable investigation.

As a tool for understanding
ÒThe purpose of computing is insight not numbersÓ wrote Richard Hamming in Numerical Methods for Scientists and Engineers in 1962. Scientific Visualization helps us to gain understanding of data, Virtual environments build on this to offer more advanced graphical environments for data investigation.

As an input into the research process and for teaching
Images are the starting point of many investigations, for example in social sciences, historical studies, medicine.

The papers presented at the conference reflected the range of uses of graphics in academic work and the tutorial sessions allowed us to address in more detail some of the underlying technologies.

Graphics as a Gateway to Information

The papers which addressed this issue were primarily concerned with access to spatial data.

David Unwin from Birkbeck College noted that maps are descriptive scale models of realities of interest which are drawn in various highly conventionalised ways. The intention is to enable us to visualise the underlying reality, but, as is illustrated by two specific map types (the choropleth and the contour), not everyone is able easily to read maps in this way. In two JISC-funded projects at Leicester University and Birkbeck College London they have created and used software tools that allow better visualisations of geographic information. In Project ARGUS, Jason Dykes used the Tcl/TK scripting language to create a Cartographic Data Viewer (cdv). This allows many simultaneous views of the same data with object linking and greatly facilitates the visualisation of area enumerated data of the type used in choroplethic maps. In the Virtual Field Course project, they have extended ideas from ARGUS to create and deliver via WWW fully navigable three dimensional terrain images for landscape visualisation. Finally, in experimental work they have used 200m resolution 1991 population density estimates for London from MIDAS to visualise the terrain provided by Greater London's population distribution in some original ways.

The KINDS project (Knowledge Based Interface to National Datasets) is linked to the MIDAS Service at Manchester Computing and is funded by JISC. Jim Petch from Manchester Metropolitan University reported that it uses the 1981 and 1991 Census, Bartholomew's Map Data and LANDSAT and SPOT image data. KINDS aims to provide a friendly interface to these national datasets and to associated software which of their very nature are complex and yet an invaluable resource. The initial strategy of KINDS was to undertake a user survey to establish a concept of user types and then assess needs as a basis for system design. A set of access and support tools were designed for creating and viewing output from single and combined datasets based on minimum user input, on providing extensive access to resources and on using a knowledge base for quality reports. KINDS is now concerned with sustainability, aiming to create a growing structured resource around the national datasets which is a repository of value-added data and other resources. This will require the development of archiving tools and extended search, help and knowledge tools. System developments also need to consider distributed data, distributed processing and open systems architecture.

As with the other themes, the conference included a tutorial session which looked in more detail at the underlying technologies. Donald Morse from the Edinburgh University Data Library presented a workshop on Putting Population on the Map. Traditionally, mapping of Census data has been a specialist activity, requiring a large investment of time and a steep learning curve. The advent over the last few years of desktop mapping / GIS packages has cut the time required and flattened the learning curve and has thus enabled the spatial dimension to be brought into the analysis and presentation of data by teachers and researchers from a wide variety of disciplines. For many, whether in the social sciences, medicine, the arts and humanities, agriculture, ecology - indeed, any discipline where the distribution of population and related resource and environmental issues are of importance - the ability to map Population Census data for analysis and presentation has assumed increasing importance. Donald Morse introduced the digitized boundary and map data and census statistics which are available to academics in the UK and how to get them. The basic thrust of the session was to emphasise how census statistics and boundaries can be combined 'on the desktop' by non-GIS specialists, teachers researchers and students, for whom geography, and in this instance, particularly boundary-based geography, can now be regarded as part of the basic toolset available to probe, manipulate, understand and present spatially-related data.

Graphics and Visualization as a Tool for Understanding

The theme of spatial data was included in this strand of the conference where computer graphics tools can help in the understanding of change. Humphrey Southall from Queen Mary and Westfield College presented the work he is leading to develop an electronic atlas of Britain. Traditional atlases are expensive. Even those which have relatively high sales volumes fall into this category and those of more academic interest often have limited, if any, colour. Yet colour is important in representation and is something which is easy to use in computer mapping. Also, even the most sumptuous full colour traditional atlas has problems presenting geographical complexity and change. The electronic atlas will use a range of formats with the markup being done using HTML. Formats from GIFs for still images to VRML for complex, dynamic, highly interactive displays will be used. The QMW team are developing the use of cartograms. Maps are created using proportional circles to display area. By animating these "new maps in new media" we can see change.

Change was a theme picked up by Mike Batty from UCL who discussed the use of GIS software on the PC to be able to examine and begin to understand the spatial distribution of urban structures, in particular the pattern of retailing. Work conducted at the UCL Centre for Advanced Spatial Analysis has used various techniques to identify and examine retail structures and to consider some of these in more detail.

Using graphics to increase understanding was addressed for a different subject area by Peter Hancock from the Department of Psychology at the University of Stirling. The team there are currently building PICS - the Psychological Image Collection at Stirling. The aim is to provide sets of images for use in psychological or vision research. Such systems may allow the recovery of general classes of images. With faces, there is also the need to be able to recover a particular individual, for example when a witness to a crime is trawling a police database. Since most suspects will not be on file, there is a need to produce a likeness, based on the witness's memory. Peter Hancock described an experimental system to help the user develop a face based on similarity of the face they are seeking to a set of randomly generated images which are constantly refined by the users judgement on the similarity to the face they seek.

The tutorial session which picked up the theme of technology to support understanding was concerned with VRML. The Virtual Reality Modelling Language has added a new dimension (the third!) to the World Wide Web. Ken Brodlie from the University of Leeds noted that VRML allows the transfer of 3D scenes rather than image over the Web. This allows the receiver to navigate and explore these worlds within a browser. Applications for teaching, research and in business are emerging with use in chemistry, urban planning, archaeology, virtual museums, shopping malls and science parks.

Using Graphical Data as an Input

Images are used as input into the research process in some disciplines such as social sciences and in medicine. Photographic collections, collections of medical images etc are often stored in higher education but are potentially fragile, often in poor condition, inappropriately stored and are relatively inaccessible. Frequently they are poorly catalogued. Digital storage gives the potential for wider access to resources with online access to images and discovery using standard systems for metadata. Marilyn Deegan of De Montfort University discussed a number of projects being undertaken within UK Higher Education. These include the ELISE (Electronic Library Service for Europe) which aims to provide access to distributed resources throughout Europe and to adopt appropriate standards for resource storage, discovery and access. Also concerned with providing a service for image access is the Knowledge Gallery which is a partnership between the commercial sector and higher education in the UK. The provision of a critical mass of content is needed if there is to be a major shift in resource provision to the electronic form and Marilyn reported on various JISC initiatives for digitization including the HELIX project funded through the Electronic Libraries programme.

A major player in the digitization of content is the SCRAN Project. The Scottish Cultural Resources Access Network is a Millennium Project, supported by funds from the UK National Lottery to build a networked multimedia resource bank for the study and celebration of Scottish material culture and human history. The word Scran in Scots means food, and to be on the scran is to be on the lookout for things to collect. Both meanings are apposite to the SCRAN project (it is gathering resources from all over the place to provide food for thought). SCRAN's funding partners are the Scottish Museums Council, the National Museums of Scotland and the Royal Commission on the Ancient and Historical Monuments of Scotland, and these have been joined on the Board of SCRAN by representatives of key organisations in the education sector. Over the five year life of the project, SCRAN expects to spend some £15 million, half from the Millennium Commission, with matching contributions in cash and in kind, including valuations of Intellectual Property Rights. In essence, in exchange for grant-aid to digitise their assets, contributors give to SCRAN a perpetual, non-exclusive, license to use them for educational purposes. Contributors retain full exploitation rights but SCRAN members can use and reuse them without further charge. The SCRAN server will be seen by most of its users as a site on the World Wide Web, but very little of the website will consist of pre-written pages. The pages that end-users will see will be created for their needs, on the fly, from the SCRAN database.

The many issues involved in the creation of digital archives of image based materials, such as SCRAN and HELIX, were discussed in a tutorial session by Jane Williams from the University of Bristol. Jane is leading the JISC Technical Advisory Support for Images service (TASI). Jane discussed issues such as: material preparation, copyright, image recording, file formats, manipulation, cataloguing and storage which need to be considered before we can look at the ways in which such resource material can be effectively accessed and integrated into electronic teaching materials and the learning process. The role of TASI, Jane explained, was to promote and advise on good practice and standards and to develop a generic framework which will accommodate different needs across higher education. Such a framework must be based on current approaches and methodologies while being flexible enough to accommodate emerging and future standards.

Some Issues

The speakers and participants were able to discuss issues within the sessions and over a very pleasant meal in the conservatory at Pollock Halls which is well located at the foot of Arthurs Seat. A number of common issues emerged from discussions and three of these are highlighted below:

the importance of usability of datasets and resources was raised by many speakers. Fitness for purpose is an issue as is the ease with which resources can be accessed and used. Graphical interfaces can play a key role
we are facing the challenge at both local and national level of offering databases of multimedia resources for both teaching and research purposes. These need to be easy to use and sustainable. We need to develop guidance and exchange views on best practice and standards if we are to offer acceptable services
a number of cross-domain issues emerged. There is a need for online resources to be searchable and accessed across traditional subject boundaries; standards for catalogues are critical here. Also, and importantly, we need to exchange ideas of techniques and tools across disciplines.

References and Contacts

Cartographic Data Visualiser (cdv)
http://www.geog.le.ac.uk/argus/

MIDAS (KINDS)
http://midas.ac.uk/

TASI
http://www.ilrt.bris.ac.uk/

VRML
http://www.sdsc.edu/vrml/

VRML Consortium
http://www.vrml.org/

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

EDINA
http://edina.ed.ac.uk/

Centre for Advanced Spatial Analysis, UCL
http://www.casa.ucl.ac.uk

QMW Project on Cartograms
http://www.qmw.ac.uk/~gbhgis

Psychological Image Collection at Stirling
http://pics.psych.stir.ac.uk/

De Montfort University International Institute for Electronic Library Research
http://www.iielr.dmu.ac.uk/

Anne Mumford (Programme Chair)
a.m.mumford@lboro.ac.uk