This paper describes our work on improving access to the content of multimodal documents containing line graphs in popular media for people with visual impairments. We provide an overview of our implemented system, including our method for recognizing and conveying the intended message of a line graph. The textual description of the graphic generated by our system is presented at the most relevant point in the document. We also describe ongoing work into obtaining additional propositions that elaborate on the intended message, and examine the potential benefits of analyzing the text and graphical content together in order to extend our system to produce summaries of entire multimodal documents.
展开▼
