Teachable machines for accessibility

45 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

ORBIT: A Real-World Few-Shot Dataset for Teachable Object Recognition (2021, background): “Other high-variation datasets include those collected by people who are blind/lowvision [18, 41, 13] (see IncluSet for a repository of accessibility datasets [19]) however, most are not appropriate for few-shot learning.”
Disability-first Dataset Creation: Lessons from Constructing a Dataset for Teachable Object Recognition with Blind and Low Vision Data Collectors (2021, background): “In terms of quality of the data, existing research developing teachable object recognisers with blind and low vision collectors [28] shows that one of the main reasons for performance degradation is the absence of the object of interest from the training examples due to challenges in photo-taking by”
Hands Holding Clues for Object Recognition in Teachable Machines (2019, background): “In the case of object recognizers with teachable machines [35, 36, 62], where machine learning models are trained and personalized by blind users, having objects partially included or even out of the camera frame is a primary factor limiting performance [36].”
Toward User-Driven Sound Recognizer Personalization with People Who Are d/Deaf or Hard of Hearing (2021, background): “However, training an ML-enabled application as a personal assistive technology can itself be inaccessible when it requires skills and abilities similar to those the application is intended to support [23, 40].”