Crowd-AI Systems for Non-Visual Information Access in the Real World

Abstract

The world is full of information, interfaces and environments that are inaccessible to blind people. When navigating indoors, blind people are often unaware of key visual information, such as posters, signs, and exit doors. When accessing specific interfaces, blind people cannot independently do so without at least first learning their layout and labeling them with sighted assistance. My work investigates interactive systems that integrates computer vision, on-demand crowdsourcing, and wearables to amplify the abilities of blind people, offering solutions for real-time environment and interface navigation. My work provides more options for blind people to access information and increases their freedom in navigating the world.