Look no text! Visual searching techniques.
Garrett, Leigh and Robinson, Amy and Collomosse, John (2012) Look no text! Visual searching techniques. In: International Plagiarism Conference, 16 - 19 July 2012, Newcastle. (Unpublished)
- Details
Over recent years there has been considerable interest, investment and engagement in the use of technology to identify sources of text based plagiarism. Despite considerable concerns in the visual arts very little work has yet been undertaken in the detection of visual plagiarism, yet students of the visual arts are required to produce numerous pieces of both visual and text based, formative and summative, assessment. Indeed the very nature of visual resources themselves makes it extremely difficult to systematically identify and detect incidents of plagiarism; instead the accusation of plagiarism is left to the knowledge and experience of academic staff, which naturally results in inconsistency of detection, approach, policies and practices.
The Spot the Difference! project (2011-2012), funded by JISC brings together the University for the Creative Arts (UCA), a specialist arts institution and major stakeholder in the detection of visual plagiarism, and the Centre for Vision Speech and Signal Processing (CVSSP) at the University of Surrey. The Visual Arts Data Service (VADS) is a research centre of UCA and holds the national image repository on behalf of the arts sector, a collection of over 120,000 items. CVSSP is the largest UK academic research centre exploring topics in image and video analysis.
The project team has developed a pilot service, iTrace, as a visual alternative to text based plagiarism detection systems such as Turnitin (Northumbria Learning) and SafeAssign (Blackboard) both of which are currently widely used to support learning and teaching in the education sector.
The pilot seeks to explore the possibility of enabling the online submission of visual artwork and seeking to identify and match those submissions to existing images on the Internet, cultural archives and previously submitted visual resources with the objective of identifying possible incidents of visual plagiarism.
No text is needed, the iTrace service makes use of the “bag of visual words” (BoVW) framework. BoVW has emerged over recent years as the most promising solution for large-scale visual search, due to its scalability and accuracy and has been shown effective in searching for books and advertising (Google Goggles), building facias (Oxford Building Search) as well as traditional CBR tasks such as colour and texture matching and related tasks such as object recognition (Sivic et al 2003, Philbin et al 2007).
By merging the principles of two quite separate technologies, including text based plagiarism systems such as Turnitin and SafeAssign with content based image retrieval projects such as DDA (Universities of Surrey and Coventry), CIRES (University of Texas) and TinEye (tineye.com) the project team has utilised CVSSP’s cutting-edge visual search technology and seek to apply it for the first time to the issue of visual plagiarism which will address the real and growing concerns of students, academic and learning support staff, and their institutions over the issue of visual plagiarism and support the development of referencing skills across in the arts.
This workshop will explore the issue of visual plagiarism; its nature and impact on learning and teaching in the visual arts; and introduce the iTrace pilot service.
Actions (login required)
 |
Edit View |