Disability-first Dataset Creation: Lessons from Constructing a Dataset for Teachable Object Recognition with Blind and Low Vision Data Collectors
By Lida Theodorou, Daniela Massiceti, Luisa M. Zintgraf, Simone Stumpf, Cecily Morrison, Edward Cutrell, M. Harris, Katja Hofmann. 2021
50 citations recorded by the source corpus.
Artificial Intelligence (AI) for accessibility is a rapidly growing area, requiring datasets that are inclusive of the disabled users that assistive technology aims to serve. We offer insights from a multi-disciplinary project that constructed a dataset for teachable object recognition with people who are blind or low vision. Teachable object recognition enables users to teach a model objects that are of interest to them, e.g., their white cane or own sunglasses, by providing example images or videos of objects. In this paper, we make the following 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 of questions to orient researchers involved in creating datasets toward reflecting on the needs of their participant community.
Subfields: Artificial Intelligence, Cognitive Neuroscience, Computer Vision and Pattern Recognition, Evaluation & Benchmarks, AI Ethics, Fairness & Society
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Mina Lee
- Katja Hofmann
- John F. Bronskill
- Qian Yang
- Luisa M. Zintgraf
- Daniela Massiceti
- Percy Liang
- Edward Cutrell
In Cognitive Neuroscience
- Cecily Morrison
- Katja Hofmann
- Percy Liang
- Edward Cutrell
- Simone Stumpf
- Lida Theodorou
- Luisa M. Zintgraf
- Daniela Massiceti
In Computer Vision and Pattern Recognition
- Cecily Morrison
- Katja Hofmann
- Percy Liang
- Edward Cutrell
- Simone Stumpf
- Lida Theodorou
- Luisa M. Zintgraf
- Daniela Massiceti
In Evaluation & Benchmarks
- Katja Hofmann
- Percy Liang
- Simone Stumpf
- Qian Yang
- John F. Bronskill
- Mina Lee
- Mohit Bansal
- Samuel R. Bowman
In AI Ethics, Fairness & Society
Citations in the corpus
- ORBIT: A Real-World Few-Shot Dataset for Teachable Object Recognition (2021, methodology): “The collection protocol was designed and validated through extensive user studies [44] and led to the key decision to capture videos rather than images of objects.”
- CoAuthor: Designing a Human-AI Collaborative Writing Dataset for Exploring Language Model Capabilities (2022, methodology): “[54] published a dataset of photos taken by blind and low vision users, for assessing how well diferent object recognition models can serve this user population.”