City Research Online

Abstract

Artificial Intelligence (AI) for accessibility is a rapidly growing area, requiring datasets that are inclusive of the disabled users thatassistive technology aims to serve. We offer insights from a multi-disciplinary project that constructed a dataset for teachable objectrecognition with people who are blind or low vision. Teachable object recognition enables users to teach a model objects that are ofinterest to them, e.g., their white cane or own sunglasses, by providing example images or videos of objects. In this paper, we make thefollowing contributions: 1) a disability-first procedure to support blind and low vision data collectors to produce good quality data,using video rather than images; 2) a validation and evolution of this procedure through a series of data collection phases and 3) a set ofquestions to orient researchers involved in creating datasets toward reflecting on the needs of their participant community.

Publication Type:	Conference or Workshop Item (Paper)
Additional Information:	© Theodorou, L., Massiceti, D., Zintgraf, L., Stumpf, S. , Morrison, C., Cutrell, E., Harris, M. T. and Hofmann, K. | ACM 2021. This is the author's version of the work. It is posted here for your personal use. Not for redistribution. The definitive Version of Record was published by ACM, http://dx.doi.org/10.1145/1122445.1122456
Publisher Keywords:	AI, accessibility, datasets, teachable object recognition, blind and low vision users
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science R Medicine > RE Ophthalmology
Departments:	School of Science & Technology > Computer Science

Export

Downloads