A tale of two images: the quest to create a story-based image indexing system
| Published date | 08 July 2014 |
| Date | 08 July 2014 |
| DOI | https://doi.org/10.1108/JD-10-2012-0130 |
| Pages | 605-621 |
| Author | Pauline Rafferty,Fawaz Albinfalah |
| Subject Matter | Library & information science,Records management & preservation,Document management |
A tale of two images: the quest
to create a story-based image
indexing system
Pauline Rafferty
Department of Information Studies, Aberystwyth University,
Aberystwyth, UK, and
Fawaz Albinfalah
Phi Research Services, Manama, Kingdom of Bahrain
Abstract
Purpose – The purpose of this conceptual paper is to consider the possibility of designing a
story-based image indexing system based on users’ descriptions of images. It reports a pilot study
which uses users’ descriptions of two images.
Design/methodology/approach – Eight interviews were undertaken to investigate storytelling in
user interpretations of the images. Following this, storytelling was explored as an indexing input
method. In all, 26 research subjects were asked to create stories about the images, whic h were then
considered in relation to conventional story elements and in relation to Hidderley and Rafferty’s (2005)
image modality model.
Findings – The results of the semi-structured interviews revealed that the majorityof interpretations
incorporatedstory elements related to setting,character, plot, literarydevices, and themes. The 52 image
stories included story elements identified in the first part of the project, and suggested that the image
modalitymodel is robust enough to deal with the “writerly”images used in this study.In addition, using
storytelling as an input method encourages the use of verbs and connotative level responses.
Originality/value – User indexing is generally b ased on paradigma tic approaches to con cept
analysis and inter pretation in the form of tagging ; the novelty of the current study is its exp loration
of syntagmatic approaches to user indexing in the form of storytelling. It is a pilot, proof of
concept study, but it is hoped that it might stimulate further interest in syntagmatic approaches
to user indexing.
Keywords Indexing, Image indexing, Image retrieval, Semantic image retrieval, Stories in images
Paper type Conce ptual paper
Introduction
The prospect of designing a story-based image retrieval system has not received
attention from scholars in the image retrieval field. Jo
¨rgensen (2003) acknowledged the
existence of story elements within some images but remarked that scholars have been
reluctant to examine this research area because “most image indexers have considered
addressing interpretive or abstract elements within an image to be beyond the scope of
their job” ( p. 251). The process of designing a story-based image indexing system
presents a challenge to indexers, not least because humans differ in their educational
and cultural backgrounds and in their perceptions, as well as in prior knowledge and
beliefs (Berger, 1972, p. 8), thus rendering any attempt to create a consistent system
for describing images difficult. Despite this problem, time and retrieval systems
have moved on since 2003, and the time might now be right, given the current interest
in user-based indexing, the social semantic web, and an emphasis on dialogic
approaches to indexing documents, to explore the potential fo r story-based indexing
of non-text documents.
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/0022-0418.htm
Received 4 October 2012
Revised 31 May 2013
18 July 2013
Accepted 28 July 2013
Journal of Documentation
Vol.70 No.4, 2014
pp. 605-621
rEmeraldGroup Publishing Limited
0022-0418
DOI 10.1108/J D-10-2012-0130
605
A tale of
two images
An earlier project reported in 2011 (Rafferty, 2011), investigated the relationship
between modality and interpretation, where modality refers to a relationship between
the sign and the real world, using a model first proposed by Hidderley and Rafferty
(2005). Four images were chosen from publicly available web sites to represent four
categories: high modality contemporary images; high modality historical images;
low modality contemporary images; low modality historical images. Distance lear ning
students from the Department of Information Studies at Aberyst wyth Universitywere
invited to assign descriptive, or denotative tags, and/or associa tive, or connotative tags
to the images. The results suggested that the Rafferty/Hidderley model was a relatively
robust categorisation model that might offer possibilities for designing ind exing
templates or information representation records, however, the high modality images
included in that study were very obviously “readerly” images. The current project adds
to the earlier work by focusing on two high modality “writerly” images. The terms
“readerly” and “writerly” derive from Barthes (1957/1973) who used them to describe
different types of texts, underpinning his textual theory with a constr uctivist view of
interpretation and meaning-making. T he “readerly” text is one in which the potential
space for reader interpretation is tightly controlled by the autho r, while the “writerly”
text is one in which the sp ace for interpretation is often actively sought in an
ambiguous text. The cu rrent project, which aims to explore stor ytelling as a basis for
image indexing, tests the tentative suggestions relating to high modality contemporary
images in the Rafferty/Hidderley model, which are that:
A non-textual information object of high modality decoded at the same historical moment,
and within the same culture and logonomic system as the encoding moment, would be
expected to evoke a limited range of denotational meanings. There may be a broader range of
connotational level interpretations. Textual anchorage could help fix the intended encoded
meaning, but decoding might evoke negotiated or oppositional interpretations. Interpretation
within a different cultural and logonomic system would potentially evoke a broader range
of connotational and denotational meanings (Hidderley and Rafferty, 2005, p. 125).
This study is partly underpinned by Schmidt and Stock’s (2009) investigation into the
possibility of creating an image indexing system based on emotions. In the course
of actualising the aim, this study critically evaluates the potential to draw on users’
interpretations of images in the design of a story-based image indexing system.
The concern is with the input side of the user-based indexing process.
Tagging offers both possibilities and drawbacks in relation to information
discovery. On the one hand, it opens up indexing to allow for multiple voices, many
approaches, efficient updating, modern terminology and language: on the other hand,
studies of tagging practice have reported that terms can be rando m, vague and
uninformative. One way of categorising tags is to distinguish between private
and public tagging, that is, tags are meant for the taggers’ benefit alone, and tags that
are assigned with a view to helping others to discover relevant information. Most web
sites that facilitate tagging allow taggers free rein about the terms they include.
This project considers whether it would be helpful to str ucture taggers’ input through
storytelling. Earlier work using a democratic indexing template designed to capture
informational tags at the level of pre-iconog raphic, iconographic and iconological
meaning suggested that a template has its uses fo r improving the exhaustivity of
indexing, however, the template was too rigid and unattractive to indexers (Hidderley
and Rafferty, 1997). Storytelling, is a pervasive and generally pleasurable form of
human communicative practice. In addition, storytelling offers a syntagmatic approach
to user-based indexing input. This project explores syntagmatic input in the form of
606
JDOC
70,4
To continue reading
Request your trial