Improving web-based human interaction proofs for blind users: implementation of real-time challenges

Return to list

This document contains no images.

Document Details
Image Details

Document Details

Title

Improving web-based human interaction proofs for blind users: implementation of real-time challenges

Identifier

DSU2014Olalere

Subjects

["Blind"]

Type

["Text"]

Genre

["Dissertations"]

Format

["pdf"]

Rights

Access to this document has been permanently and indefinitely restricted by the author until or unless the author submits a written request to Special Collections & Archives to change the access restrictions and/or make it freely available. Only Albert S. Cook Library staff members may access this file for administrative purposes unless written permission from the author is presented to staff members of the Towson University Special Collections & Archives department. Copyright protected, all rights reserved.

General Notes

Doctoral thesis research directed by the Department of Computer and Information Sciences.

Extent

xiii, 220 pages

Language

["English"]

Repository

["Towson University. Albert S. Cook Library. Special Collections and University Archives"]

Collection Name

["Towson University Graduate Theses and Dissertations"]

Author

Olalere, Abiodun;

Degree Date

2014

Department or Program

["Towson University. Department of Computer and Information Sciences"]

Abstract

This research is focused on efforts towards improving Human Interaction Proofs (HIPs) - making HIPs more secure and usable for people with visual impairments. Web forms are protected from automated Bots through the use of Completely Automated Public Turing Test to tell Computers and Humans Apart (CAPTCHAs). However, most CAPTCHAs available are visual, which makes them unusable by people with visual impairments. A few audio CAPTCHAs are available, but are difficult to use, and often considered unsecure.

This research explored the design of a new audio CAPTCHA called the SoundsRight Audio CAPTCHA (simply called SoundsRight or SoundsRight CAPTCHA). The SoundsRight Audio CAPTCHA works by presenting to users a series of generic sounds (dogs barking, bells ringing, drums beating, etc.). Users must indicate each time they hear a specific type of sound by pressing the spacebar on the keyboard. If all instances of a sound category have been correctly identified, the user is allowed to access the computing service. If not, the user is directed to take the CAPTCHA test again.

The research was executed in three stages respectively, the design and usability study of the SoundsRight Audio CAPTCHA, introduction of sound masking to the SoundsRight Audio CAPTCHA and real-life implementation of the SoundsRight Audio CAPTCHA. Controlled studies documented that SoundsRight is very usable with success rate above 90%. The masking study shows that the introduction of sound masking and different depths of masking have significant impact on success rate and blind users are better at solving audio CAPTCHAs than sighted users. It was also determined that success rates under real-life implementation differ from laboratory studies. Given the real-time nature of SoundsRight and the high success rates recorded by users, the SoundsRight Audio CAPTCHA is expected to be a good alternative to existing audio CAPTCHAs. Findings in this research provide insight into the design and the factors that would be helpful to consider in taking SoundsRight to the mainstream of wider adoption and usage on the World Wide Web.

Contents

Design of the SoundsRight Audio CAPTCHA -- Usability study of the SoundsRight Audio CAPTCHA -- Masking study design -- Usability evaluation of masking on the SoundsRight Audio CAPTCHA -- Design of SoundsRight for real-life implementation -- Real-life implementation results

Thesis Committee Members

["Feng, Jinjuan Heidi","Chakraborty, Suranjan","Chakraborty, Joyram","Brooks, Tim"]

Date Published

09 September 2014

Author (Controlled)

["Olalere, Abiodun M."]

Thesis Advisor

["Lazar, Jonathan"]

Degree Note

(D. Sc.) -- Towson University, 2014.

Return to list