Automated tactile graphics translation: in the field

  • Authors:
  • Chandrika Jayant;Matt Renzelmann;Dana Wen;Satria Krisnandi;Richard Ladner;Dan Comden

  • Affiliations:
  • University of Washington;University of Washington;University of Washington;University of Washington;University of Washington;University of Washington

  • Venue:
  • Proceedings of the 9th international ACM SIGACCESS conference on Computers and accessibility
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

We address the practical problem of automating the process of translating figures from mathematics, science, and engineering textbooks to a tactile form suitable for blind students. The Tactile Graphics Assistant (TGA) and accompanying workflow is described. Components of the TGA that identify text and replace it with Braille use machine learning, computational geometry, and optimization algorithms. We followed through with the ideas in our 2005 paper by creating a more detailed workflow, translating actual images, and analyzing the translation time. Our experience in translating more than 2,300 figures from 4 textbooks demonstrates that figures can be translated in ten minutes or less of human time on average. We describe our experience with training tactile graphics specialists to use the new TGA technology.