Transforming diagrams’ semantics to text for visually impaired

Authors: Cross, C., Cetinkaya, D. and Dogan, H.

Journal: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Volume: 12200 LNCS

Pages: 339-350

eISSN: 1611-3349

ISBN: 9783030497125

ISSN: 0302-9743

DOI: 10.1007/978-3-030-49713-2_24


Using models and diagrams is a very useful and effective tool for representing information and systems in a graphical form to communicate and understand them better. On the other hand, graphical representations bring extra cognitive load and the process for understanding the diagrams is long and tedious in most cases for the visually impaired. To solve this problem, semantics of the diagrams should be converted to a different format that is both human and machine readable as well as communicable for the visually impaired. Most existing diagramming tools are not easily usable for the visually impaired as a tool for creating and using diagrams. In this paper, we propose an online system for defining specific diagrams and converting their semantics to text which can have a speech output for the visually impaired. We present analysis and design of this online system as well as a proof of concept prototype implementation. The prototype system provides create, save, load and transform features and tested with participants to recreate the diagrams using the automatically generated text output. Our case study showed that the results are very promising and the proposed solution can provide a way to correctly and accurately represent the information in diagrams textually.

Source: Scopus

Transforming Diagrams' Semantics to Text for Visually Impaired

Authors: Cross, C., Cetinkaya, D. and Dogan, H.

Conference: HCI International 2020

Dates: 19-24 July 2020

Source: Manual

Transforming Diagrams' Semantics to Text for Visually Impaired.

Authors: Cross, C., Cetinkaya, D. and Dogan, H.

Editors: Marcus, A. and Rosenzweig, E.

Journal: HCI (20)

Volume: 12200

Pages: 339-350

Publisher: Springer

ISBN: 978-3-030-49712-5

Source: DBLP