Design of and Subjective Response to On-body Input for People with Visual Impairments
Uran Oh, Leah Findlater · 2014 · Proceedings of the 16th International ACM SIGACCESS Conference on Computers & Accessibility (ASSETS) · doi:10.1145/2661334.2661376
Summary
This paper investigates on-body input — using the skin as an input surface — as an alternative or complement to touchscreen interaction for people with visual impairments. Since blind users do not need the visual display of a mobile device, on-body input offers potential advantages: an always-available input surface, increased tactile and proprioceptive feedback compared to a smooth touchscreen, and the ability to interact without pulling out a phone. The researchers conducted a lab study with 12 visually impaired participants (9 totally blind, 3 with low vision; mean age 44.3) involving two tasks. In Task 1, participants created gestures for five mobile actions (previous, next, open, home screen, e-mail) at five body locations: same hand, other hand-palm, other hand-back, forearm, and face/neck. They then rated each location for ease of use, physical comfort, and social acceptability across various contexts (places, audiences, physical constraints). In Task 2, participants performed basic VoiceOver tasks using four input conditions in a 2x2 design: phone vs. on-hand input, one vs. two hands. The on-hand sensing system used a lightweight ring with a colour marker tracked by camera and a capacitive touch sensor, translating hand gestures into VoiceOver commands (double tap, swipes, long tap) sent via Bluetooth to an iPhone 4S.
Key findings
The study revealed that social acceptability was the dominant factor in location preferences, overriding ease of use and physical comfort. The other hand-palm was the overall most preferred location (8 of 12 participants), with acceptance rates of 95.8% across contexts. Face and neck was overwhelmingly the least preferred (acceptance rate 35.4-48.6%), despite ease of use and comfort ratings being higher than same hand in some measures. Participants explained that face/neck gestures would appear rude, attract unwanted attention, or interfere with conversations. The forearm had moderate ratings but participants worried others might think they "might have fleas." For gesture creation, directional swipe was the most common gesture across all locations. Participants used four strategies to create distinct gestures: varying the basic gesture type, number of fingers, specific body landmarks (e.g., pointing to palm vs. fingertip), and which fingers were used. In the phone vs. on-body comparison (Task 2), two-handed phone input was most preferred (8 participants), but one-handed on-body input was considered the most versatile — 10-11 of 12 participants would use it regardless of physical constraints (seated, standing, walking, one hand busy). Phone input was easier with two hands while on-body was easier with one hand, creating a clear complementary relationship. All 12 participants noted difficulty using their phone while walking. One-handed on-body input was particularly valued for situations where one hand holds a cane or guide dog leash — a common constraint for blind mobile users that two participants specifically raised. Participants' preferences shifted between tasks: largely positive about on-body input in the imagined scenarios of Task 1, but more nuanced after using a working system in Task 2, highlighting the importance of evaluating with functional prototypes.
Relevance
This is the first in-depth study of on-body input designed specifically for people with visual impairments, and its findings have implications for the growing field of wearable and gesture-based interaction. For accessibility practitioners and designers of mobile assistive technology, several insights stand out. First, the physical constraint of holding a cane or guide dog leash is a defining use case for blind mobile users that is rarely considered in mainstream mobile design — on-body input directly addresses this by freeing one hand. Second, the dominance of social acceptability over practical factors challenges the assumption that users will adopt whatever is most efficient; blind users, like all users, care deeply about how their interactions appear to others, and this concern should be central to wearable and gesture design. Third, the complementary relationship between phone and on-body input suggests that on-body interaction is most valuable not as a replacement for touchscreens but as an additional input channel for specific contexts (walking, one hand busy, quick interactions without pulling out the phone). The design recommendations — favour hand locations, use familiar touchscreen gesture types, support body landmarks for gesture differentiation, and design for one-handed use — provide concrete guidance for future accessible wearable input systems.
Tags: mobile accessibility · on-body interaction · gesture interaction · blindness · wearable technology · eyes-free interaction · social acceptability · input methods