Title: Making Visual Language Accessible: A Crowdsourced Approach

Advisors: Richard Ladner

Supervisory Committe: Richard Ladner (Chair), Jaime Snyder (GSR, iSchool), Meredith June Morris, Alan Borning, and Katharina Reinecke

Abstract: This talk presents research to make visual language more accessible to sign language users and low-vision readers. Worldwide, there are about 70 million deaf people using a sign language as their first language, and almost 300 million people with visual impairments. However, much of society and the technical world communicates through written text. This excludes many people from full participation, as sign languages lack a standard written form, and low-vision readers struggle to access text visually.

In this talk, I will demonstrate how to design, build, and evaluate systems that include deaf and low-vision people in our text-based communication platforms by providing equal access to visual language. These solutions are powered by data-driven approaches, but it can be challenging to collect sufficient data from minority disabled populations. Our methods overcome data scarcity challenges by making use of other populations, such as student language learners and crowdsourcing workers. In this talk, I will present three projects that embody this approach: 1) a crowdsourced American Sign Language (ASL) dictionary, 2) smartfonts and livefonts, alternate scripts that improve legibility for low-vision English readers, and 3) the first animated ASL character system.

CSE 303
Monday, May 1, 2017 - 10:00 to 12:00