Blind Photographers and VizSnap: A Long-Term Study
Dustin Adams, Sri Kurniawan, Cynthia Herrera, Veronica Kang, Natalie Friedman · 2016 · Proceedings of the 18th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '16) · doi:10.1145/2982142.2982169
Summary
This paper reports on a two-month longitudinal study with 13 blind participants (ages 18-65, nine totally blind and two with some light perception, seven adventitiously and six congenitally blind) using VizSnap, a free iPhone app designed to help blind people organize and browse a photo library without sight. VizSnap addresses a key challenge: while blind people actively take photographs, the images themselves contain only visual information, making it difficult to browse, identify, and retrieve specific photos later. VizSnap automatically records ambient audio while the user is aiming the camera, captures time, date, and GPS location metadata, and allows the user to record an optional voice memo — all with fewer gestures than the default iPhone camera app with VoiceOver. Participants met with researchers every two weeks for questionnaires and photo retrieval exercises. The researchers also analyzed thousands of forum posts from a Flickr group called "Blind Photographers" to understand common photography challenges, finding that over a third of photos had no obvious quality problems, while the two biggest issues were photos being cut off or out of focus. The study addressed three research questions: what kinds of photos blind photographers take, whether audio and metadata assist with memory retrieval, and what issues arise during long-term use.
Key findings
Participants took a total of 800 photographs across the study, categorized into 11 types: outdoor scenery (225, the largest category), individual person (103), food/drink (82), vehicle (73), toy/craft (69), plant (59), household item (54), animal/pet (46), group (46), electronics (40), and whole room (36). For photo retrieval, participants were described five photos and asked to locate them using only audio, time, date, and location metadata. In Session 2, 49 out of 54 photos (90.7%) were correctly identified by the 11 participants who completed the task. In Session 3, accuracy dropped to 85% (34/40), likely because participants now had twice as many photos to search through. Both ambient audio and voice memo were consistently rated as the most helpful metadata for photo retrieval — in Session 2, ambient audio was rated most useful by 54% and voice memo by 38%; in Session 3, voice memo was rated most useful by 55% and ambient audio by 22%. Date and time were the next most useful, while location was rarely cited as most helpful. Participants' confidence in sharing photos grew over the study: Facebook posting increased from 8% in Session 2 to 44% in Session 3, and email sharing rose from 23% to 33%, suggesting that regular use of VizSnap builds confidence in photo quality. Common problems included difficulty with gestures (4 reports), not knowing photo quality (2), and audio volume being too low (4). Most requested features were photo quality feedback, camera flash access, and louder volume.
Relevance
This research demonstrates that photography is a meaningful creative and social activity for blind people, challenging assumptions that visual media is irrelevant to non-visual users. The 800-photo corpus reveals that blind photographers engage with the same range of subjects as sighted photographers — people, food, nature, pets, vehicles — underscoring that photographs serve as memory anchors and social currency regardless of whether the photographer can see them. For accessibility practitioners, the key design insight is that non-visual metadata (ambient audio, voice memos, timestamps, location) can make visual media navigable without vision, achieving 85-91% retrieval accuracy. The finding that ambient audio — an automatically captured, effortless form of metadata — was rated as or more useful than deliberate voice memos suggests that passive context capture is a powerful accessibility strategy. The growing willingness to share photos publicly over the study period highlights that confidence, not capability, is often the barrier to blind people's participation in visual social media — and that tools providing assurance about content quality can address this barrier.
Tags: blindness · photography · mobile accessibility · social media · iPhone · VoiceOver · user research · image accessibility