College of Computing and Digital Media Dissertations

Date of Award

Winter 11-13-2013

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

First Advisor

Dr. Rosalee Wolfe

Abstract

Internet technologies have expanded rapidly over the past two decades, making information of all sorts more readily available. Not only are they more cost-effective than traditional media, these new media have contributed to quality and convenience. However, proliferation of video and audio media on the internet creates an inadvertent disadvantage for deaf Internet users. Despite technological and legislative milestones in recent decades in making television and movies more accessible, there has been little progress with online access. A major obstacle to providing captions for internet media is the high cost of captioning and transcribing services.

To respond to this problem, a possible solution lies in automatic speech recognition (ASR). This research investigates possible solutions to Web accessibility through utilization of ASR technologies. It surveys previous studies that employ visualization and ASR to determine their effectiveness in the context of deaf accessibility. Since there was no existing literature indicating the area of greatest need, a preliminary study identified an application that would serve as a case study for applying and evaluating speech visualization technology. A total of 20 deaf and hard-of-hearing participants were interviewed via video phone and their responses in American Sign Language were transcribed to English.

The most common theme was concern over a lack of accessibility for online news. The second study evaluated different presentation strategies for making online news videos more accessible. A total of 95 participants viewed four different caption styles. Each style was presented on different news stories with control for content level and delivery. In addition to pre-test and post-test questionnaires, both performance and preference measures were conducted.

Results from the study offer emphatic support for the hypothesis that captioning the online videos makes the Internet more accessible to the deaf users. Furthermore, the findings lend strong evidence to the idea of utilizing automatic captions to make videos comprehensible to the deaf viewers at a fraction of the cost. The color-coded captions that used highlighting to reflect the accuracy ratings were found neither to be beneficial nor detrimental; however, when asked directly about the benefit of color-coding there was support for the concept. Further development and research will be necessary to find the appropriate solution.

Share

COinS