A Wearable Input Mechanism for Blind Users of Computers Based on Mental Mapping of Thumb-To-Phalanx Distances
Yash Naik · 2019 · Proceedings of the 21st International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2019) · doi:10.1145/3308561.3356109
Summary
This student research abstract presents Mentalitype, a glove-based wearable keyboard designed to provide blind users with an affordable and portable text input mechanism for computer-based tasks. The system maps 32 push-buttons to the phalanges (finger bones) of both hands, with each finger segment assigned a character. The thumb remains uncovered and is used to identify and press the buttons on the other fingers, leveraging the assumption that blind users have a stronger mental model of their own body positions than of external devices like touchscreens or standard keyboards. The device uses an Arduino Due microcontroller for USB interfacing and works with the NVDA screen reader to provide audio feedback for each keystroke. The design includes 26 alphabetic keys arranged across the phalanges, plus 6 modifier keys (power, number/alphabet mode toggle, spacebar, enter, caps lock, and control). The system supports both alphabet and number input modes, with buttons reconfigurable to suit individual user preferences. The research is motivated by the limitations of existing accessible input methods: touchscreen-based solutions like BrailleTouch and BrailleSketch require battery-operated devices, standard QWERTY keyboards present spatial learning challenges, refreshable braille displays cost between and ,500, speech-to-text systems are unreliable in noisy environments, and chording gloves require memorizing complex key combinations.
Key findings
A user study with 5 blind participants across 5 sessions (each up to 20 minutes) demonstrated promising results. Participants averaged an entry rate of 3.57 words per minute (WPM), outperforming BrailleType's initial rate of 1.45 WPM. Two participants reached 6.0 WPM by the final session, suggesting that extended practice could approach the Chording Glove's initial 8.9 WPM. The average character error rate (CER) was 3.58%, substantially better than BrailleType (8.91%) and BrailleSketch (10.6%), and competitive with Gaines' touchscreen method (2.08%). The average backspaces per tap (BPT) was 0.03, lower than Gaines' 0.068, indicating high input accuracy. Both CER and BPT showed a decreasing trend across sessions, suggesting that accuracy improves with continued practice. The study used words from the Oxford 3000 list and progressively increased task complexity from individual letters to full sentences of up to 25 characters.
Relevance
This research addresses a genuine gap in accessible input technology by proposing a low-cost, body-referenced alternative to expensive braille displays and unreliable speech recognition. The concept of mapping input to proprioceptive body awareness rather than external spatial layouts is a creative approach that could inform future assistive device design. However, the study is limited by its small sample size (5 participants) and short practice duration (100 minutes total). The 3.57 WPM average speed remains far below standard keyboard rates, raising questions about practical daily use. The authors note planned future work with autistic children and users with symbrachydactyly, which could broaden the device's applicability. For accessibility practitioners, this work highlights that affordable hardware solutions using commodity components like Arduino boards can deliver competitive accuracy, even if speed remains a challenge to be addressed through extended training and iterative design refinement.
Tags: blindness · wearable technology · text input · assistive technology · alternative keyboards · haptic feedback