Analysis of Use Need Image Archives

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

The effect of postings information on searching behaviour

Analysis of user need in image


archives

Linda H. Armitage and Peter G.B. Enser visual images. The synthesis of these processes is
generating rapidly developing capabilities in the realm
University of Brighton, UK
of the electronic image and operates with accelerating
force on our ability to gain physical access to our
Received 14 November 1996 visual heritage.
However, our ability to gain logical or subject access
to that visual heritage has not enjoyed comparable
advances. In part, this is due to the well-recognised
Abstract.
problems which beset the indexer of still and moving
This paper describes a project in which an analysis was image material [1–13]. As with any information
undertaken of user queries addressed to seven libraries retrieval function, the indexing problem is condi-
which manage archives of widely varying still and moving tioned by the characteristics of user need for the
image material. The sampling procedure is described, in source material.
which queries obtained from each library were broadly
Surprisingly, user need for visual imagery has been
categorised by image content, identification and accessi-
the subject of very little enquiry. Keistler [14] has
bility. Attention is focused on the image content requests,
for which a categorisation based on facet analysis is devel- studied queries addressed to the photographic collec-
oped. The analytical tool which is used for this purpose is tion of the National Library of Medicine and Ornager
based on a schema already well established for the analysis [15] has described journalists’ requests submitted to a
of levels of meaning in images. number of newspaper archives. A systematic repre-
The project demonstrates the possibility of formulating a sentation of information need in the visual domain,
general categorisation of requests which seek widely incorporating an in-depth analysis of some 2,700
different still and moving image material. The paper requests serviced by a major, general-purpose picture
concludes with observations on the potential value of archive has been reported by Enser and Mcgregor [16]
embedding such a schema within the user interface of and Enser [17]. The project reported here, in extending
unmediated-query visual information retrieval systems.
the latter analysis, sought to develop a general-purpose
categorisation of user requests for still and moving
visual images. In generating a greater understanding
1. Introduction of user need for such knowledge records, it sought
also to contribute towards a better-informed view of
There is widespread recognition that we live in an effective retrieval techniques for visually encoded
increasingly visually-oriented society; one in which information.
textual representation no longer has to be the domi-
nant paradigm for information and instruction. This
paradigm shift has been reinforced by major techno- 2. Data collection
logical thrusts which have been applied to the capture,
storage, compression, transmission and display of
2.1. Participating archives

Correspondence to: Dr P.G.B. Enser, School of Information Seven picture libraries participated in the study. Two of
Management, University of Brighton, Watts Building, the archives, the National Film and Television Archive
Moulescomb, Brighton BN2 4GJ, UK. Tel: +44 1273 600900. (NFTVA), maintained by the British Film Institute
Fax: +44 1273 642405. E-mail: P.G.B.ENSER@brighton.ac.uk (BFI), and the BBC Natural History Unit (NHU),

Journal of Information Science, 23 (4) 1997, pp. 287–299 287


Analysis of user need in image archives

maintain collections of moving imagery. The former Table 1


supports a wide and general user base interested in the Library sample sizes
world of film and television, while the latter serves a
much narrower range of ‘expert’ users interested in the Library No. of image content
queries
specific subject domain of natural history.
Of the five libraries concerned with still imagery, the The Witt Library, Courtauld
focus of Glasgow’s Mitchell Library (Glasgow Room) Institute 186
and of Birmingham Central Library (Local Studies Wellcome Institute for the
Division) is local history in different forms, mainly History of Medicine 232
photographic. The Mitchell Library also hosts the The Mitchell Library, Glasgow 170
North British Locomotive Company’s photographic Birmingham Central Library 294
archive. The collections at Birmingham cover a BBC NHU 301
BFI NFTVA 365
geographical range which is much more extensive than
NMR Air Photographs Division 201
the local area, because they contain a number of inter-
Total 1,749
nationally important collections by renowned photog-
raphers. Both libraries cater for a broad user-base.
The Air Photographs Division of the National
Monuments Record (NMR), maintained by the Royal unmediated queries were taken from their collections
Commission on the Historical Monuments of England, of letters and/or faxes sent in by users.
deals with one particular form of imagery: namely, Initial analysis led to the identification of the
aerial photographs. This national collection covers the following major types of query:
whole of England, users being drawn from a wide (1) image content, i.e. find me some images of . . .;
base, but especially from professions such as town (2) identification/attribution/provenance checking;
planning and engineering. (3) accessibility of image/artist(s) work, e.g. owner-
The Witt Library, Courtauld Institute of Art, ship/viewing availability;
University of London, and the Wellcome Institute for (4) miscellaneous, e.g. requests for administrative
the History of Medicine include among their holdings procedures only or unusable queries.
substantial collections of images relating to art history. Only the first of these was of direct concern, and the
The Wellcome concentrates on the subject domain of sample sizes of image content queries obtained from
medicine, whereas the Witt covers art history in the participating libraries is shown in Table 1. Extracts
general. Both serve academic communities within from each sample are reproduced in the Appendix.
their wide user-bases.
In summary, the seven participating picture
libraries/archives embrace both general and specialist 3. Analysis of image content requests
collections of still and moving images and a variety of
types of user. In aggregate, together with the then
3.1. Broad categorisation
Hulton Deutsch Collection Ltd (now the Hulton Getty
Collection Ltd) employed in the earlier study, they A similar categorisation to that adopted in the earlier
provide a comprehensive coverage of visual image Hulton study [16, 17], where user requests were found
types and client needs. Further details about the image to be easily represented in terms of a 2 ´ 2 matrix of
holdings of each archive may be found in Armitage unique/non-unique, refined/unrefined queries, could
and Enser [18]. be applied to much of the sample data in the project
reported here. Unique subjects are defined as those
concerned with named individuals, one-off events,
2.2. Query collection
singular objects or locations. Examples of library
In consultation with library staff, query sample selec- sample requests for unique and non-unique subjects
tion was guided by the intention to collect approxi- are shown in Tables 2 and 3.
mately 200 records from each site, representative of However, some of the library samples (the Witt
the usual flow of requests for visually-oriented infor- Library in particular) contained requests which sought
mation within that library or section. The recording of images by a certain artist or creator rather than of a
the requests implies a process of mediation. In the case certain subject. Such queries clearly formed a separate
of the two public libraries, however, samples of category. Requests for known items, e.g. photograph

288 Journal of Information Science, 23 (4) 1997, pp. 287–299


LINDA H. ARMITAGE AND PETER G.B. ENSER

Table 2 album of Wellcome Tropical Laboratories, Khartoum,


Examples of unique subject queries or Whistler’s ‘Flying Rocket’, were also differentiated
from other types of image content query and allocated
Unique subject query Library to a separate category. In the case of the BBC NHU and
Adam and Eve Witt BFI NFTVA collections, requests by programme name
Pasteur Wellcome for an entire programme’s footage, e.g. a good 35mm
photograph of Sir R. Balfour Mitchell print of ‘Oh What a Lovely War’ also fell into this latter
Billy Matchett in costume Birmingham category.
Rust the Fallow Deer BBC NHU The distribution of the query collections over the
Alan Turing BFI NFTVA resulting four categories is revealed in Table 4.
Hogarth Club (1880s) Witt Consideration of this data reveals that the ‘By . . .
Colney Hatch Asylum Wellcome [artist]’ category is most significant in the Witt Library
HMS Volunteer Mitchell
sample. This is to be expected, given the subject
Yardley Brick Works – Red Hill Birmingham
domain of this collection: art history assigns a funda-
St Helena BBC NHU
earliest footage of St Paul’s BFI NFTVA mental importance to the individual producer of the
RAF Syerston, 1940–1960 NMR artefact. Whereas, while the Wellcome works with
Supper at Emmaus Witt many traditional art historical forms, its primary
Sir Henry Thompson operating on purpose is to supply imagery on the history of medi-
Napoleon III Wellcome cine. The uniformity across the other three categories
July 3rd 1951 Judy Garland . . . at for these two libraries, however, is striking.
Birmingham Hippodrome Birmingham The slightly higher value of the ‘By . . . [artist]’ cate-
Japanese Expo 1996 BBC NHU gory for the BFI NFTVA also expresses the difference
Janet Street-Porter’s speech at last between film as an art form which is known by, and
yr’s Edin . . . + TV fest[ival] BFI NFTVA
forms part of, an individual artist/director’s œuvre,
and film produced for the BBC NHU which does not
place primary emphasis on the individual producer.
Table 3 With this pair of libraries, there is uniformity in
Examples of non-unique subject queries known-item searching, but significant differences
between unique and non-unique subject searching.
Non-unique subject query Library The subject domain of natural history provides little
twins Witt in the form of unique subjects, since flora and fauna
Japanese people Wellcome are regarded as groups of homogeneous items. It is not
portraits of ministers Mitchell usual practice to name or locate individual examples
coal miners (c. 1840–c.1890) Birmingham of plant or animal in precise settings. Thus, for the
starving puffin. Puffin. 440.606.920 BBC NHU BBC NHU, it is not surprising that there is a marked
women jockeys BFI NFTVA
kitchens Witt
lying-in hospitals Wellcome Table 4
Clyde cruisers Mitchell Image content query category breakdown by library
shopping arcades Birmingham
high-tech aquariums BBC NHU Library No. of Image content category
liners 1920s BFI NFTVA queries
By Known Unique Non-
gambling – cards and dice Witt named items subjects unique
electrotherapy Wellcome artist subjects
illustration of archery Mitchell
air raids Birmingham Witt 186 10.8% 10.3% 31.4% 47.6%
running emu BBC NHU Wellcome 232 1.7% 12.0% 34.3% 51.9%
archaeology in Mexico BFI NFTVA Mitchell 170 0.0% 24.1% 50.0% 25.9%
Birm’ham 294 1.0% 43.5% 44.2% 11.2%
BBC NHU 301 0.0% 42.3% 1.7% 55.9%
BFI NFTVA 365 2.7% 40.3% 24.7% 32.3%
NMR 201 0.0% 13.9% 86.1% 0.0%
Total 1,749

Journal of Information Science, 23 (4) 1997, pp. 287–299 289


Analysis of user need in image archives

Table 5
Panofsky-Shatford mode/facet matrix

Iconography Pre-iconography Iconology


(Specifics) (Generics) (Abstracts)

Who? individually named kind of person or thing mythical or fictitious


person, group, thing being
(S1) (G1) (A1)
What? individually named kind of event, action, emotion or
event, action condition abstraction
(S2) (G2) (A2)
Where? individually named kind of place: place symbolised
geographical location geographical,
architectural
(S3) (G3) (A3)
When? linear time: cyclical time: season, emotion, abstraction
date or period time of day symbolised by time
(S4) (G4) (A4)

emphasis on non-unique subject searching compared imagery on carnivals clearly addresses a non-unique
with the BFI NFTVA. event. A request for images of the Rio Carnival, 1986,
In contrast, the NMR sample is dominated by unique however, certainly relates to a unique time, place and
over non-unique subject searching. This is attributable event. A request for imagery on Rio carnivals straddles
to custom and practice in the library, which insists on the two categories uneasily.
users defining a geographical region within which In the Hulton study, the anomaly posed by subject
searches are to be made. As such, every query is ring- matter which contained elements of both uniqueness
fenced into a singular location. and non-uniqueness was resolved by the concept of
In the case of the Mitchell and Birmingham pair, ‘refinement’ [16, 17]. Thus, Rio carnivals would have
requests for locally-based imagery necessarily seek been categorised as a request for a non-unique subject
unique locations. However, the almost equal domi- (carnivals) refined by location (Rio). In essence, the
nance of known-item searching in the Birmingham process of adding refiners is an application of facet
sample is accounted for by queries which address their analysis. The high incidence of such queries in the
nationally important collections by renowned photog- current investigation encouraged the authors to adopt
raphers, which transcend local history. a more formal facet-based analytic technique in order
The apparent simplicity and functionality of the to determine if a robust, functional representation
four-element categorisation of queries described above could be achieved, while also revealing broad unifying
masks the complex nature of the relationship between characteristics across the range of project libraries.
unique and non-unique subjects in the visual domain.
At the root of the problem is the fact that, in the
3.2. Categorisation based on mode/facet analysis
natural order of physical existence, an entity can
always be interpreted into an hierarchy of related The problem was approached by adopting Panofsky’s
super-concepts and sub-concepts; as one navigates a modes of image analysis [19], as explicated by Markey
path through such a concept hierarchy, it may not be [9] and further refined by Shatford [5]. The analytical
obvious at what level one encounters the property of tool which was derived took the form of the template
uniqueness. This problem was manifest in queries shown in Table 5, which is adapted from Shatford.
which appeared to invoke subject matter which was By representing each cell of the matrix as a code
either unique or non-unique, depending upon the (shown in brackets), and returning to the ‘carnivals’
vantage point from which the relevant concept hier- example given earlier, single and multi-faceted
archy was being viewed. For example, a request for requests, including those which incorporate uneasy

290 Journal of Information Science, 23 (4) 1997, pp. 287–299


LINDA H. ARMITAGE AND PETER G.B. ENSER

Fig. 1. Image facet content data cumulative percentage chart – mode/facet analysis.

combinations of unique and non-unique subject faceted queries. The percentage sub-totals and total per
matter, can be represented in a straightforward manner library are given as an indication of the level of
as mode/facet combinations: overlap between categories.
● carnivals: G2 Fig. 1 gives the cumulative percentage spread across
● Rio carnivals: S3 + G2 the seven project libraries, illustrating the general
● the Rio Carnival, 1986: S2 + S4. similarity of pattern of query form.
Image content data from all the library samples was As can be seen from the diagram, the mode/facet
processed using the matrix as a tool of analysis. Table analysis reveals both the broad unifying characteris-
6 gives selected examples. tics of the pairs of libraries and the different emphases
Table 7 shows the overall categorisation of the between pairs, as discussed earlier. In addition, it
project data which was achieved as a result of mode/ gives a clearer picture of any individual library’s char-
facet analysis. Percentage totals per library add up to acteristics. For example, the BBC NHU queries are
more than 100% because of the incidence of multi- strongly characterised by requests which ask for

Journal of Information Science, 23 (4) 1997, pp. 287–299 291


Analysis of user need in image archives

Table 6
Examples of queries categorised by mode/facet analysis

Queries Specifics Generics Abstracts


S1 S2 S3 S4 G1 G2 G3 G4 A1 A2 A3 A4

Mahomed, the Brighton Shampooist *


QED: ‘300 Million Year War’ *
St Catherine’s Dispute with the Philosophers * *
ref. to Petre family and Ingalstone Hall, Essex * *
Napoleon at Jaffa, 1799 * * *
paintings and illust’s to Goldsmith’s ‘Deserted Village’ 1770–1860 * *
Keats’ teachers * *
LP [Living Planet series]: ‘Open Ocean’ – waves, dolphin leap, beluga * * *
Hickman experimenting with anaesthesia on animals * *
Iconography of St Anthony Abbot * *
Martyrdom of St Lawrence *
Ranelagh Gardens *
Shots of moon *
Images of New Amsterdam, 17th C * *
Illustrations of menagerie at Stowe Landscape Gardens, 19th C * * *
Writers in Hollywood * *
people at Ascot, Windsor and Henley – being entertained * *
Game reserves in Kenya, Tanzania * *
Oxford Univ. [19]40s graduation ceremonies * * *
1920s Hungarian peasants * *
20th C illustration for article on adultery * *
Victorian asylums * *
Image of Janus’ head in 18th C British [school] * *
Allegories depicting Memory, 20th century image desirable * *
3-D film audiences in the 50s * * *
Suffragettes *
Corsetry * *
Soldiers talking clearly * *
Killing for a living: 1. lion – savannah, 2. army ants – jungle, 3. . . . * * *
painting of [any] woman in kitchen * *
Port[ugese] Man O’War on beach with people * *
Toads at night * *
Arthurian legend in British art * *
people who are against or phobic about new technology * *
Iconography of Dominican saints and Blesseds * *
Sea birds rising up to sky from seascapes at dawn + twilight * * * *
[Animal] nest-building *
Chinese foot binding *
[people] working in nuclear power station + wearing protective . . . * *
mist and torrential downpour onto jungle * *
Physiological laboratory *
Underwater, no human – 20–30" *
Equinox *
Vampires, spirits *
Depiction of Vanity *
S1 S2 S3 S4 G1 G2 G3 G4 A1 A2 A3 A4

292 Journal of Information Science, 23 (4) 1997, pp. 287–299


LINDA H. ARMITAGE AND PETER G.B. ENSER

Table 7
Sample percentages of mode/facet categorisation by library

Witt Wellcome Mitchell Birm’ham BBC BFI NMR


NHU NFTVA

Specifics:
S1: who? ind. named person, group,
thing 45.5 44.9 33.9 39.9 58.3 64.4 13.9
S2: what? ind. named event, action 3.4 4.2 0.6 2.5 1.7 9.0 1.5
S3: where? ind. named geographical
location 10.2 8.9 50.6 54.0 11.6 8.7 87.1
S4: when? linear time: date, period 20.5 5.1 13.1 24.5 0.4 8.3 17.5
Sub totals 79.6 63.1 98.2 120.9 72.0 90.4 120.0

Generics:
G1: who? kind of person, thing 45.5 22.9 46.4 14.7 54.5 14.4 0.5
G2: what? kind of event, action,
condition 18.8 39.3 4.8 7.4 30.2 15.1 0
G3: where? kind of place: geographical,
architectural 4.5 5.1 15.5 1.2 13.6 1.6 36.3
G4: when? cyclical time: season,
time of day 0 0.5 0 0 1.7 0 0

Sub totals 68.8 67.8 66.7 23.3 100.0 31.1 36.8


Abstracts:
A1: who?: mythical/fictitious being 10.2 2.8 0 0 0.4 0 0
A2: what? emotion, abstraction 6.3 0.9 0 0 1.2 0.3 0
A3: where? place symbolised 0 0 0 0 0 0 0
S4: when? emotion, abstraction
symbolised by time 0 0 0 0 0 0 0

Sub totals 16.5 3.7 0 0 1.6 0.3 0


Totals 164.9 134.6 164.9 144.2 173.6 121.8 155.8

generic groups of animals, but which also supply events as to need no extra detail. An S2 code in these
suggested named titles of programmes either for cases seems appropriate.
content extraction or as examples of similarity of In contrast, the Queen’s Christmas Day message
proposed coverage. The significance of the specifics (1988) might be seen as an event in itself, but, because
group in the Birmingham sample reflects the concen- it is a repeated event, it must be differentiated by a
tration on genealogical tracing, which works by clari- particular year’s broadcast. In this case, an S1 + S4
fying a particular link in a chain at a time. Its coding might be more suitable. However, the query
significance in the NMR sample reflects the nature of cheetahs running on a greyhound course in Haringey
the individualised query process. in 1932 satisfies specific location and time facets, but
The generally low incidence of category S2 among deals with a generic group of animals involved in
the library samples warrants some comment. It might some general activity. Clearly though, the request
be argued that an individually named event can be relates to a one-off event and differs markedly, because
described adequately with some combination of S1: of its startling nature, from queries asking for dogs
named person(s)/thing(s) + S3: location + S4: linear running, of which there may be many such events at
time – but, for example, the query Churchill’s funeral the same place within the same time span. An S2
or a request such as the Moon landing mention neither coding seems more fitting than a cumbersome G1 + G2
place nor time, but represent sufficiently memorable + S3 + S4. Thus, although the S2 category is not well

Journal of Information Science, 23 (4) 1997, pp. 287–299 293


Analysis of user need in image archives

represented among the query samples, there does seem Another general problem of any model is the need
to be a need to cater for unique events. This is espe- sometimes to incorporate attributes to do with the
cially so in the case of the BFI NFTVA, which gener- medium of the imagery. For example, images of a
ated a significant number of queries such as Bombing well-known modern personality may include paint-
of Guernica, 1937 and landspeed record of ’56. ings, drawings, etchings, photographs, film, cartoons,
A more general problem is posed by queries which collages and so on. These might be regarded as sub-
specify geographical location. Any named location can selections (G1) within a general request for images of
be regarded as unique (and therefore specific) simply that personality (S1). However, a request for images of,
because it occupies different space to any other named say, a gallery’s exhibition catalogue is quite different
location. However, there is a significant difference from images of the exhibition itself or the gallery. In
between a query such as scenes of London and St this sense, the catalogue form (G1) drives the query,
John’s Cottage, Monument Lane, Edgbaston. The latter over and above the specific information on the event
describes an entity which is distinguishable from its (S2) or place (S3). This is a particular feature of the
surroundings and which could be placed easily within Witt Library data. Thus, object attributes can become
the S3 category. The former begs many more questions a necessary part of the subject description.
before it can become a meaningful query, although The low incidence of assignments to the Abstracts
it is concerned with a named location. As it stands, it categories is a partial reflection of the nature of the
may represent, for example, a request for the sort of libraries involved in the study. All seven (together
cityscape skyline found in many cities; in which case, with the Hulton library in the earlier study) are picture
the query aligns more with the generics G3 category. or film archives, as distinct from stock-shot libraries.
Alternatively, it may be asking for the sort of tourist The emphasis in the latter lies with the provision of
attractions that do distinguish this city from another, a rapid-response service to requests for images, where
whereupon it could be placed in the S3 category. It it is frequently the case that the client is an advertising
may, of course, represent any number of other inter- agency and the request addresses the iconologic
pretations. content of an image. Among stock-shot libraries, there
The following two queries illustrate a less stark is little tradition of indexing or classifying their collec-
contrast, but generate the same problem. Images of the tions, or recording the queries addressed to them.
Birchfield area of Birmingham in the 1930s is equiva- Accordingly, such organisations were not included in
lent to the ‘London scenes’ request, in that easily the study reported here. The fact that the Witt Library
definable boundaries are lacking in terms of geog- sample demonstrates the highest concentration of A-
raphy, architecture or topic, even though the area category facets is understandable, given the nature of
asked for may be quite small. Similarly, the finite the Witt’s collection and user base.
boundaries of Home Farm, Thorneaugh, Cambs may
be visually indistinguishable from any neighbouring
farm’s fields, especially if aerial photography is the 4. Conclusion
medium in question. Both these queries ask for indi-
vidually named locations, but could involve a variety The project reported here demonstrated that there are
of different types of image. observable similarities in image query formulation
The problem is one of defining the scale and size of across a range of different libraries and that, despite
what a single geographical unit should encompass and variations in library rationale, stock and user profile,
therefore any overarching geographical framework a general characterisation of queries can be formu-
within which queries could be graded as specific or lated.
generic. Different library environments develop ways The analytical framework which supports the query
of handling the geographical element according to the characterisation reported here is rooted in a schema
kind of stock they hold and the sort of queries already well established for the analysis of image
addressed. For example, within the project libraries, content. The dual applicability of the schema to the
the BFI NFTVA indexes countries, cities and towns. characterisation both of images and the queries which
The NMR ensures definable limits by a combination address those images offers some encouragement to
of grid reference and scale of the aerial photography. the view that the embedding of such a schema within
For the purposes of consistency within the project the user interface might offer a pathway towards the
data, the decision was taken to assign an S3 code to effective processing of unmediated transactions within
any named location. a visual information retrieval system.

294 Journal of Information Science, 23 (4) 1997, pp. 287–299


LINDA H. ARMITAGE AND PETER G.B. ENSER

The incidence of such unmediated transactions is Victorian paintings on, or incorporating the subject of,
increasing under the influence of those advances in ‘Invention’.
information and communications technology which Romney portrait of C.F. Greville. 1781.
have brought CD products and networked delivery of
Bernard Samual Morles – images of foundlings in
digitised images to the marketplace. Some of the
1870s.
libraries which participated in the study reported here,
for example, while continuing to offer a traditional Don Juan in French and English painting (19th
user interface in which the human intermediary plays century) as figure in literature/opera (Mozart, Hoffman,
a crucial role in receiving and processing queries, Byron, Shaw, etc) and figure/symbol of the hero.
already provide, or are engaged in the development of, Representation of the Muse ‘Clio’ or allegorical repre-
such image delivery mechanisms. The project reported sentation of ‘History’, 17th and 18th C Images of
in this paper, in seeking to further our understanding Death, Grief, Mourning – 19th C British.
about visual information need, aims to contribute
A2 Sample of queries collected from the Wellcome
towards our developing perceptions of good interface
Library
design for such visual information retrieval systems.
Urology.
Corsetry.
Appendix Industrial health.
Synthetic antimalarial agents.
A1 Sample of queries collected from the Witt
Remmelin anatomy.
Library
Smallpox and fevers.
Gambling – cards and dice.
video: Battle for the Mind (WIHM/IC/70).
‘Atlas’.
Deans of St Thomas’ Hospital Medical School.
. . . images of Native Americans or others murdering
colonists’ children especially babies, in the New Typhus in England and Ireland.
World, e.g. Hannah Durton’s children . . . Material is x-ray.
needed for article on infanticide with ref to Cooper’s Bethlem, Ashe’s and St George’s hospitals.
‘Last of the Mohicans’.
women in medicine.
Any aspect of lavender cultivation – plants, lavender
Childbirth in history, midwifery, obstetrics.
fields, selling, etc.
Sir Henry Thompson operating on Napoleon III.
Depictions of Lucy from ‘The Bride of Lamermoor’ or
Lucinde Lamermoor. Hermaphrodite, giants.
Reynolds painting of Wm Boyce mentioned in the Nelson lying in state.
DNB and ‘Portraits of Musicians Exhibition’ at South history of adolescence.
Kensington Museum 1885 handwritten catalogue by plague of Marseilles, late 18th cent. anatomical prac-
Sir George Scharf. tices.
The depiction of vanity in painting, the depiction of Sir Joseph Fayrer (Indian Medical service).
the female figure looking in the mirror, etc.
Victorian asylums.
Railways.
A3 Sample of queries collected from the Mitchell
Ann Wharton – wife of Thomas, Lord Wharton
Library
1659–1680.
Dance halls – photos.
Likenesses of Capt. John Money and other balloonists
+ balloon accidents. Scottish Daily Express – 1951 illustration.
Lawyers + legal subjects. Green’s Playhouse – dancing.
Weapons in use. Copeland Rd (Govan) area – photos.
Portraits of Harriet Houner. HMS Volunteer – photo.
Illustrations of menagerie in Stowe landscape Gardens Sir R. Balfour – photo.
19th C. 1938 Exhibition – posters.

Journal of Information Science, 23 (4) 1997, pp. 287–299 295


Analysis of user need in image archives

Theatre Royal photocopies. Specific interior sections. Sir Benjamin Stone – photograph of two mutes (1901)
Clyde cruisers – photos. Sir Benjamin Stone – photograph of the funeral of a
Balshagary School – photo. Birmingham shopkeeper (1901).
Horsedrawn fire engine, ambulance, etc – photos. . . . 35mm slides of the following:
Clyde Dock 1925/26 strikers – print. a) the frontispiece of Charles Pye’s Birmingham
Glasgow Port 18th century – photo. Directory of 1797 (. . . photocopy attached from the
Belgian steam locos. Birm. Arch. Soc. Transactions 1937 . . .)
Irish immigration to Glasgow – photos of weavers’ b) a portrait of Matthew Boulton.
social conditions 18th, 19th century. . . . either a postcard or photograph of Aston Manor
‘A Venetian fish market’ – painting. . . . [believed to be] HQ of the 8th Battalion Royal
Warwickshire Regiment, formed in 1908. . . .
Sketch plan of Glasgow 1547.
First moving pictures in Glasgow. . . . 1. Church Street, Hereford, between 1875 and 1930
2. Colwall, Herefordshire, particularly a property
Portraits of ministers.
called Colwall Mill, Old Colwall in the 1860s or 1870s
NBL: sketch of a brass plate from a locomotive; 3. Any photo of a blacksmith at Colwall in the 1870s,
specific, info. on train. 1880s or 1890s
Queen Street. Specific street number, illustrations of. 4. Any of Welland or Castlemorton, Worcestershire
A4 Sample of queries collected from the 5. Any of the Verger of Hereford Cathedral between
Birmingham Central Library 1875 & 1930.
. . . enquiries about the Francis Frith photo collection . . . one or two good pictures of the latest Bull Ring
. . . to obtain old photos of Denver, in Norfolk . . . to Shopping Centre, plus confirmation that some part of
research how it [Manor Farm] formerly appeared to aid this lies on or near an earlier Market Cross and/or
a planning application, as it is now a listed building Market House fro Mediaeval times? . . . [plus] any
. . . [or] any photos of Denver. picture postcards of Shopping Arcades . . . around the
turn of the century. . . .
. . . the picture of Sir Charles Rivers Wilson (a relative,
1831–1916) in ‘Vanity Fair’ 9 Nov 1878 by Pellegrini. . . . early photographs of the area around St Mary’s
Chapel, Whittall Street, including St Mary’s Row, Love-
. . . research is the history of Japanese photography in day Street, and a street that once existed before con-
the UK collections whether there is [sic] any Japanese struction of the General Hospital, ‘Weaman Row’. . . .
photographs especially before 1945, in your collec-
tions. . . . . . . whether the Collection [Benjamin Stone
Collection] contains any photographs that may exist of
. . . [re] Charge of the Light Brigade at Balaclava 1853 street parties taken at the time. . . .
. . . [to] trace a photograph of Corporal Martin
. . . [re] the 50th anniversary of VE day . . . any
Langfield, a bugler in the above charge, . . . depicted
photographs that may exist of street parties taken at
in one of the several books written on the battle.
the time . . . .
. . . if you have any negatives of scenes in North Curry,
. . . photographs of the roof garden of Lewis’s depart-
Somerset, Witham Friary, Somerset, Alfriston, Sussex
ment store as it was in the 1950s and 1960s. . . .
or Wyke Regis, Dorset.
. . . Aston Grammar School . . . any prints of the school
. . . i) . . . any photographs of motoring subjects taken
at this time [i.e. 100 years ago]?. . . .
by F.R. Logan (successor to Blyth Clayton) of Cannon
Passage, Birmingham. He photographed motoring . . . research into Royal burial places in Europe . . . [so
subjects extensively, including photos of trials, speed request for] photographs of the Royal vaults under the
trials, hill-climbs etc outside the city . . . Albert Memorial Chapel, St George Chapel, Windsor . . . .
ii) motoring photographs, 1900–1930, by any other . . . [to] look in the index for the Frith Collection to
photographers. see if there are any pictures of windmills in the
. . . a photograph of Mr Hugh Gilzean-Reid – MP in following Leicestershire villages: Eaton, Hoton,
1885 for Aston Manor. Sewstern, Ab Kettleby, Seagrave, Foxton. . . .
. . . one photograph of each of the following works . . . to acquire a photographic copy of St Judas Church
from your collection: (sepia finish) for family research files. . . .

296 Journal of Information Science, 23 (4) 1997, pp. 287–299


LINDA H. ARMITAGE AND PETER G.B. ENSER

A5 Sample of queries collected from the BBC NHU MOTHS – ESPECIALLY IN THE SAME SHOT. 12]
Library BOGS + CREATURES RELATED TO BOGS [DRAGON-
UNDER WATER NO HUMANS. 20–30". FLIES, ETC] – NOT ‘LIVING ISLES’! WOO NIGHT LIFE
– GET WATERWALKERS. TICKET TO THE WILD.
BATS. BUTTERFLIES. TIME-LAPSE – SENSE OF
TIMING. ALL TIGERS. DOLPHINS SWIMMING. GALAPAGOS
IS. MONTAGED. LIVE ACTION. STUNNING SHOTS.
JAPANESE EXPO – 1996. CITIES.
KILLER WHALE. GOLDEN EAGLE. KUMEDO
CLOSE SHOTS OF CRABS BREEDING. UK ONLY, DRAGON. ENTER THE DRAGONS.
MED SHOTS OKAY TOO [TO CUT INTO FOOTAGE
SHOTS OF MOON.
OF LINCOLNSHIRE COAST FOR ITEM ON AREA
IMPORTANT FOR CRABS, UNDER THREAT. A6 Sample of queries collected from the BFI
OIL FROM UNDERNEATH – BIG OIL. BELUGA FILM. NFTVA Library
LP OPEN OCEAN – WAVES, DOLPHIN LEAP, 1937 television recording of Rothschild sale.
BELUGA.
BBC Healthwatch last night on microwave hazards of
KOMODO DRAGONS, SEXUAL ENCOUNTERS OF mobile phones.
THE FLORAL KIND. LP: WORLD’S APART. TNW:
On the Waterfront for open air screening o/s RFH
ENTER THE DRAGONS.
[Royal Festival Hall].
KILLING FOR A LIVING. 1] LION SAVANNAH. 2]
Urban life.
ARMY ANTS – JUNGLE. 3] PINNEPEDS. 4] KILLER
WHALES. Frida: Naturaleza Vita [Mexican film].
WOLVES FEEDING. 421.310.11. old footage for the Lincolnshire Life Museum.
SURVIVAL PROG ON MONGOOSES. TOGETHER Liners 1920s.
THEY STAND 22866. Late Show Special on produc[tion] of Coriolanus.
NATURE – TURKEY – GENERAL SHOTS. [for entertainer who impersonates Queen Victoria]
ARMADILLOS. Queen Victoria and Romanov dynasty.
BIRTH OF EUROPE 1. Liverpudlians/Liverpool in film and TV.
BLUEBOTTLE. ANY OTHER FLYING BUZZING ‘Right the First Time’ training video.
INSECT. Cheetahs running on a greyhound course in Haringey
SWARM TRIMS – PORT MAN O’WAR ON BEACH/ in 1932.
WITH PEOPLE Interferon [the drug].
VHS RONI POSSUM – A NEW ZEALAND NIGHT- Doc on the rebuilding of Glyndebourne 22/5/94.
MARE [WILD SOUTH]\PLAYING POSSUM – ABC. films of D Day, H. Bogart, C. Gable.
ANY OTHER POSSUM PROGS. people who are against or phobic about new tech-
EIDER DUCK ALL KILLED. [1] NEED ––––- HORIZON nology.
FILM – HAVE IT. [2] 4 Spp of CEM. [3] oystercatchers. man/women working in nuclear power stations +
FROM ‘THE NATURAL WORLD – SKYHUNTERS’ wearing protective clothing.
ARE THERE ANY TRIMS OF TURKEY VULTURES/ Queen’s Christmas Day message (1988).
BLACK VULTURES CIRCLING OVER A CORPSE?
Rejuvenation treatments; cryonics; immortality.
PLEASE FIND ME THE FOLLOWING CLOSE UP/BIG
Earliest footage of St. Paul’s.
CLOSE UPS MID CLOSE UPS IN A BRITISH OR
BRITISH LOOKING ENVIRONMENT: SECRET John Hurt + Miranda Richardson.
NATIVE LYMN. 1] BADGER BEING ELUSIVE [NOT
BROCKSIDE OR LIVING ISLES]. 2] WOOD MICE. 3] A7 Sample of queries collected from the NMR Library
BANK VOLE – IF YOU CAN GET THESE ITEMS ONLY Note: each request assumes the form: initials of
SELECT TO (sic) MORE EXAMPLES OF researcher/year/month/query number within month.
ITEMS 4 TO 8. 4] WEASEL. 5] ROE DEER. 6] STOAT. Place name or feature: map sheet number, grid reference
7] MOLES – WOO UNEARTHING THE MOLE. 8] (+ expanded search area). The specialist collection refers
SHREW. 9] BRITISH BUTTERFLIES. 10] BATS. 11] to the collection of oblique aerial photographs.

Journal of Information Science, 23 (4) 1997, pp. 287–299 297


Analysis of user need in image archives

PJR/95/4/2. Dates: All. Scales: 1:12,001. Collection/no. searches:


Tregarrick Tor (in stereo): 201 SX/240/710, Vertical/1
244/714 FEM/95/4/13. 128/065
Dates: All. Scales: < 15,000. Collection/no. searches: 148/057
Specialist/1, Vertical/1 Giantlees Farm, Northumberland: 81 NU/151/049
FEM/95/4/3. 141/041
‘CROPMARKS AT BAYSTON HILL –- 1969 117/050
. . . NMR SJ4809/1 –- WAB137/90 . . . illustrates ‘Three Dates: All. Scales: < 1:22,001. Collection/no. searches:
small sub-circular cropmarks.’ [oblique print SH Specialist/1 Vertical/1
4809/1 frame 90 film WAB 137/90] CLK/95/4/14 see (user) list (i.e. . . . and continues . . .)
FEM/95/4/4. Place name Nat. grid ref. OS map
Brean & Berrow, Somerset: 300/590 Plessey NZ 240 800 081 Felton NU 175 003 081
182 ST/290/510 Choppington NZ 252 838 081 Alnwick NU 155 165 081
Dates: 1946/7. Collection/no. searches: Vertical/1 Ogle NZ 124 797 088 Swinhoe NU 203 264 075
Hebron NZ 203 907 081 Warrenford NU 130 286 075
FEM/95/4/5.
N Witton NZ 099 910 081 Sionside NU 090 337 075
Near Willingdon: 199 TQ/575/022 (+ 500 m)
Stobswood NZ 225 942 081 Goatscrag NT 984 368 075
Dates: All. Scales: < 1:15,001. Collection/no. searches:
Acklington NU 208 017 081 Berry Hill NT 945 407 075
Specialist/1 Vertical/1 Aller Dean NT 990 471 075
FEM/95/4/6. Dates: 1940 – 46. Scales: < 1:25,001. Collection/no.
Thetrundle: 197 SU/877/110 (+ 500 m) searches: Vertical/15
Dates: All. Scales: < 1:15,001. Collection/no. searches: CLK/95/4/15.
Specialist/Vertical/1 Mirfield: 104 SE/180/190
FEM/95/4/7. 110 SE/220/230
Warwick Castle: 151 SP/283/647 (+ 400 m) Dates: All. Scales: < 1:15,001. Collection/no. searches:
Dates: All. Scales: All. Collection/no. searches: Specialist/1 Vertical/1
Specialist/1 [per site] CLK/95/4/16.
FEM/95/4/8. North Lincolnshire Long Barrows [Project] [see sample
1. Neston 173 ST/865/685 (+500 m) from 9 coversearch req. forms]
2. Corsham 173 ST/870/705 (+ 500 m) [1.] Broughton, S. Humbs: 112 SE/970/080 SW corner
SE/980/090 NE corner
Dates: pre-1950. Scales: All. Collection/no. searches:
Dates: All. Scales: > 1:15,000. Collection/no. searches:
Vertical/1 [per site]
Specialist/1 Vertical/1 [i.e. per req. form]
FEM/95/4/9.
CLK/95/4/17.
Fyfod Project: 173 SU/105/670,
Fernhurst, W. Sussex: 186 SU/89/28
170/745
91/29
Dates: All. Scales < 1:15,001. Collection/no. searches:
Dates: < 1945. Scales: All. Collection/no. searches:
Specialist/1 Vertical/1 Specialist/1 Vertical/1
FEM/95/4/10. CLK/95/4/18.
4 quarter sheets. Notts, NMP: 1. SE/70/SW Hesketh Arms, Rufford, W. Lancs: 108 SD/4606/1555
2. SE/79/NW, SW Dates: All. Scales: < 15,001. Collection/no. searches:
3. SK/59/SE Specialist/1 Vertical/1
Dates: All. Collection/no. searches: Specialist/3
CLK/95/4/20.
FEM/95/4/11. Bromsgrove: 139 SO/94/69
Manchester: 109 SJ/84/93 (km') /96/71
Dates: All. Scales: 1:12,001. Collection/no. searches: Dates: All. Scales: < 22,001.
Specialist/1 Vertical/1 DS/95/4/21.
FEM/95/4/12. Manor Farm, Great Shefford: 174 SU/380/755 + 750?
Lower Wheelton: 103 SD/60/21 3 ´ km' 380/753
/59/21 Dates: 1940–56. Collection/no. searches: Specialist/1
/59/20 Vertical/1

298 Journal of Information Science, 23 (4) 1997, pp. 287–299


LINDA H. ARMITAGE AND PETER G.B. ENSER

Acknowledgements [9] K. Markey, Computer-assisted construction of a


thematic catalog of primary and secondary subject
Financial support for this project reported in this matter, Visual Resources III (1983) 16–49.
[10] M.G. Krause, Intellectual problems of indexing picture
paper was provided by the British Library Research
collections, Audiovisual Librarian 14(2) (1988) 73–81.
and Innovation Centre (Project no. RDD/G/235), whose [11] M. Hogan, C. Jorgensen and P. Jorgensen, The visual
help is gratefully acknowledged. thesaurus in a hypermedia environment: a preliminary
We would also like to record our thanks to the staff exploration of conceptual issues and applications. In:
in the seven participating libraries who assisted and D. Bearman (ed.), Hypermedia and Interactivity in
encouraged us during the course of the investigation. Museums: Proceedings of an International Conference,
The libraries concerned are identified with their Pittsburgh, PA, October 1991 (Technical Report 14)
permission. (Archives and Museum Informatics, Pittsburgh, PA,
Thanks are due also to Laura Baird, for her help in 1991), pp. 202–221.
data collection in the early stages of the project. [12] B.C. O’Connor, Representation and the utility of
moving image documents, Annual Conference of ASIS
(1986) 237–243.
[13] R. Hidderley et al., Capturing iconology: a study in
References retrieval modelling and image indexing. In: M. Collier
and K. Arnold (eds), Electronic Library and Visual
[1] A.E. Cawkell, Indexing Collections of Electronic Information research. ELVIRA 2. The Proceedings of the
Images: A Review (British Library Research Review 15) Second ELVIRA Conference, held in May 1995 at De
(British Library, London, 1993). Montfort University, Milton Keynes (Aslib, London,
[2] A.E. Cawkell, Picture-queries and picture databases, 1995), pp. 79–91.
Journal of Information Science 19 (6) (1993) 409–423. [14] L.H. Keistler, User types and queries: impact on image
[3] P.G.B. Enser, Pictorial information retrieval (Progress in access systems. In: R. Fidel et al. (eds), Challenges in
Documentation), Journal of Documentation 51(2) (1995) Indexing Electronic Text and Images (ASIS Monograph
126–170. Series) (Learned Information Inc, Medford, NJ, 1994)
[4] P.G.B. Enser, An indexing-free approach to the retrieval pp. 7–22.
of still images. In: A. McEnery (ed.), Proceedings of the [15] S. Ornager, The newspaper image database: empirical
13th British Computer Society Information Retrieval supported analysis of users’ typology and word associ-
Specialist Group Research Colloquium, University of ation clusters. In: Proceedings of the 18th Annual
Lancaster, 8–9 April 1991 (British Computer Society, International ACM SIGIR Conference on Research and
London, 1991), pp. 41–55. Development in Information Retrieval, 9–13 July,
[5] S. Shatford, Analyzing the subject of a picture: a Seattle, WA, pp. 212–218.
theoretical approach, Cataloguing and Classification [16] P.G.B. Enser and C.G. Mcgregor, Analysis of Visual
Quarterly 5(3) (1986) 39–61. Information Retrieval Queries (British Library R&D
[6] S. Shatford, Describing a picture: a thousand words are Report 6104) (British Library, London, 1992).
seldom cost effective, Cataloging and Classification [17] P.G.B. Enser, Query analysis in a visual information
Quarterly 4(4) (1984) 13–30. retrieval context, Journal of Document and Text
[7] S. Shatford Layne, Some issues in the indexing of Management 1(1) (1993) 25–52.
images, Journal of the American Society for Information [18] L.H. Armitage and P.G.B. Enser, Information Need in the
Science 45(8) (1994) 583–588. Visual Document Domain (British Library Research and
[8] E. Svenonius, Access to nonbook materials: the limits Innovation Report 27) (British Library, London, 1996).
of subject indexing for visual and aural languages, [19] E. Panofsky, Studies in Iconology: Humanistic Themes
Journal of the American Society for Information in the Art of Renaissance (Harper and Row, New York,
Science 45(8) (1994) 600–606. 1962).

Journal of Information Science, 23 (4) 1997, pp. 287–299 299

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy