A Face Recognition Application for People with Visual Impairments: Understanding Use Beyond the Lab
summary
A Facebook Messenger bot called Accessibility Bot on Facebook Messenger runs on Android with TalkBack, uses Facebook’s face‑recognition API and a model tested at 97% lab accuracy vs perceived 0.2–0.9 in the field, and was deployed to six people with visual impairments for a seven‑day diary study. Four of six find it helpful, common failures are poor photos and aiming the camera, the double‑tap trigger to protect bystanders slows use, and unexpected uses include selfies and photographing printed photos.
AI-written summary · about
chapters (9)
- 0:00Research motivation
- 1:41System design goals
- 3:24Prototype functionality
- 6:28User study methodology
- 7:56Key study findings
- 10:52Social and safety impact
- 12:49Reflection and discussion
- 14:31Conclusion and acknowledgments
- 15:50Q&A session
related talks
22mAndroid Accessibility: Understanding Blind Users to Meet Their Needs
09mImageAlly: A Human-AI Hybrid Approach to Support Blind People in Detecting...
20mScene text access: a comparison of mobile OCR modalities for blind users
37mSeeing is Believing: Mobile Vision API Deep Dive
45mBe My Eyes - Leveraging Generative AI to help the blind and visually impaired people
20mWe Need to Talk About Accessibility in Chatbots
39mEasy Android Accessibility
49mThe impact of AI on people with disabilities