Object Artefact Script Cayless
From ESIWiki
Object, Artefact and Script (October 8-9)
img2xml: Linking Image and Text with SVG
Hugh Cayless (New York University)
Annotation and linking (or referring) have been described as “scholarly primitives,” basic methods used in scholarly research and publication of all kinds. This presentation will deal with ongoing work to prototype a method for deep-linking images of texts (particularly handwritten texts that are not amenable to OCR) to transcriptions and annotations using vector graphic tracings of the writing. Given a Scaleable Vector Graphic (SVG) representation of a written text, one can use standard techniques to link transcriptions and annotations to the image at the level of letter, word, line, etc. In addition, the vector representation may make certain types of structure detection simpler than techniques which utilize raster images. The letter and word shapes encoded in the SVG may be extracted and displayed on their own, as well as in context.
The theoretical, as well as the practical applications of this technique are interesting. It holds out the possibility of exploiting the presentational logic of the text (i.e. the effects the layout and rendering of the text have on its meaning) in a digital way. It also may allow scholars to capture more information in the course of creating a transcription (which necessarily involves some filtering and interpretation of what the scholar reads on the support) than is currently possible.
