VEMI Lab

Published: 2013

Publication Name:

Abstract:

Hari Prasath Palani. Unpublished Master’s Thesis, Nov. 2013, UMaine. (N.A. Giudice: thesis advisor)

Accessing graphical material such as graphs, figures, maps, and images is a major challenge for blind and visually impaired people. The traditional approaches that have addressed this issue have been plagued with various shortcomings (such as use of unintuitive sensory translation rules, prohibitive costs and limited portability), all hindering progress in reaching the blind and visually-impaired users. This thesis addresses aspects of these shortcomings, by designing and experimentally evaluating an intuitive approach —called a vibro-audio interface— for non-visual access to graphical material. The approach is based on commercially available touch-based devices (such as smartphones and tablets) where hand and finger movements over the display provide position and orientation cues by synchronously triggering vibration patterns, speech output and auditory cues, whenever an on-screen visual element is touched. Three human behavioral studies (Exp 1, 2, and 3) assessed usability of the vibro-audio interface by investigating whether its use leads to development of an accurate spatial representation of the graphical information being conveyed. Results demonstrated efficacy of the interface and importantly, showed that performance was functionally equivalent with that found using traditional hardcopy tactile graphics, which are the gold standard of non-visual graphical learning.

One limitation of this approach is the limited screen real estate of commercial touch-
screen devices. This means large and deep format graphics (e.g., maps) will not fit within the screen. Panning and zooming operations are traditional techniques to deal with this challenge but, performing these operations without vision (i.e., using touch) represents several computational challenges relating both to cognitive constraints of the user and technological constraints of the interface. To address these issues, two human behavioral experiments were conducted, that assessed the influence of panning (Exp 4) and zooming (Exp 5) operations in non-visual learning of graphical material and its related human factors. Results from experiments 4 and 5 indicated that the incorporation of panning and zooming operations enhances the non-visual learning process and leads to development of more accurate spatial representation. Together, this thesis demonstrates that the proposed approach —using a vibro-audio interface is a viable multimodal solution for presenting dynamic graphical information to blind and visually-impaired persons and supporting development of accurate spatial representations of otherwise inaccessible graphical materials.