City Research Online

Selecting and tailoring of images for online news content: a mixed-methods investigation of the needs and behaviour of image users in online journalism

Frankowska-Takhari, S. (2018). Selecting and tailoring of images for online news content: a mixed-methods investigation of the needs and behaviour of image users in online journalism. (Unpublished Doctoral thesis, City, University of London)

Abstract

This mixed-methods investigation explores how image professionals in online journalism search for, select and use images from large online collections. Further, findings from this exploration are used to devise and evaluate a needs-based practical solution for improvement to image retrieval.

The exploratory stage included semi-structured interviews and observations in situ and provided several important contributions to the current understanding of the needs and behaviour of image users in fully disintermediated environment of the online newsroom. This study found that these image users are creative professionals and self-taught, yet, confident image searchers. When illustrating news content, they apply a shared knowledge of how a specific image function (e.g., dominant image) must be presented visually to reach its full communication potential. This common understanding of image communicative functions has two implications on how these professionals search for and select images. Firstly, they begin searches with clear image needs pre-defined on multiple levels of image description, including visual image features, and their behaviour is consistent with targeted searching. This contradicts previously reported preference for browsing as the typical mode of searching in online image collections. Secondly, they do not easily compromise on image needs related to visual features. When searches prove ineffective, they resort to editing skills and tailor the available images to match their original needs.

Further, it was found that the choice of images for headline content can in fact be predicted by a set of 11 visual image features. The features were extracted from a collection of artefacts created in the observation sessions and described by means of the Visual Social Semiotics (VSS) framework. The feature set was implemented as a filtering mechanism in a prototype and evaluated in a within-subjects experimental design study with image professionals. This experiment showed a significant positive change in the behaviour of users when interacting with images pre-filtered strictly to their visual needs, not observed in the baseline system. This was demonstrated through users’ ability to immediately engage in the inspection of images on a level of detail, and to make straightforward selections. Images from the experimental sets required no or only minimal tailoring as confirmed in the final VSS-based survey with independent image experts.

Other important contributions of this investigation include the updated models. Firstly, the illustration task process framework, originally proposed in Markkula and Sormunen (2000), has been refined to include the image tailoring phase where creative professionals apply editorial treatment before publication. Further, the observations revealed that verifying of images, consistent with the feature in Ellis et al.’s model (Ellis et al., 1993), was an activity critical to making selection decision in online journalism. Therefore, Conniss et al.’s model of the image searching process (Conniss et al., 2000) has been updated to include the verifying phase.

The investigation concludes that in order to meet the needs of creative image professionals in online journalism, image retrieval systems must support targeted searching, and facilitate direct access to required images that can be easily verified for authenticity. The proposed multi-feature filtering system firmly rooted in the image users’ needs, appears to be a step towards automating image retrieval.

Publication Type: Thesis (Doctoral)
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
T Technology > TR Photography
Z Bibliography. Library Science. Information Resources > ZA Information resources > ZA4050 Electronic information resources
Departments: Doctoral Theses
Doctoral Theses > School of Mathematics, Computer Science and Engineering Doctoral Theses
School of Mathematics, Computer Science & Engineering > Computer Science
School of Mathematics, Computer Science & Engineering > Computer Science > Human Computer Interaction Design
URI: http://openaccess.city.ac.uk/id/eprint/21947
[img]
Preview
Text - Accepted Version
Download (8MB) | Preview

Export

Downloads

Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login