Towards Semantically Enhanced Audio Rendering of Equations

Abstract

At present STEM education beyond middle school is largely inaccessible to visually impaired students in countries like India. Access to equations, tables, charts and figures are the key bottleneck. With the increased penetration of screen reading software, effective audio rendering of equations can significantly help in making many of the e-texts accessible. Unfortunately, linear syntactic rendering of equations can create considerable cognitive load even for relatively simple equations. In this paper we propose an architecture to extract contextual semantic of equations based on the local definitions. This will help in adapting audio rendering of equations based on their contextual semantics.