The Thematic Illustrator: An Automatic Illustrative Approach to Enhancing Narrative Cohesion
Charlie Hargood, David E. Millard, Mark J. Weal · ePrints Soton (University of Southampton) · 2011
Adaptive or generated narratives offer personalised and dynamic delivery of narrative content to users. However their generation or adaption is often based on the literal content of the narrative and not the subtext ignoring many important narrative principles such as narrative cohesion. Based on our existing work on a thematic model we present the Thematic Illustrator; a system capable of generating content relevant themed illustrations for short stories. The system aims to emphasise core themes and as a result improve the thematic cohesion of the story. We investigate whether this has a tangible effect on the cohesion of the narrative as a whole by using five variables to measure the cohesion of a story with themed and unthemed illustrations. Our results show that thematic noise can be reduced and that there is a demonstrated link between thematic cohesion and some other aspects of narrative cohesion.