Automatic Video-to-Audiotactile Conversion of Golf Broadcasting on A Refreshable Pin Array

Abstract

Video accessibility is an important but challenging research question. In this study, we implemented and evaluated a system that converts video content into audio clips and tactile icons without losing context using a refreshable pin array display. The suggested system converts contextual information of the video to audio description and tactile scenes, allowing users to hear and touch. As an initial target, we selected golf broadcasting, a popular sport in the BLVI community, which has a clear context yet relies heavily on visual features and provides limited information through audio. We extracted contextual information through computer vision to deliver information such as scores and the trajectories and results of shots. Then, we converted them to audio via Text-to-Speech and tactile icons on the pin array. We evaluated the system by conducting a perception experiment and a usability survey, and the results showed that the system effectively converted the information. © 2024 Owner/Author.