A Portable Device for the Translation of Braille to Text
Iain Murray, Andrew Pasquale · 2006 · Proceedings of the 8th International ACM SIGACCESS Conference on Computers and Accessibility (Assets '06) · doi:10.1145/1168987.1169030
Summary
This paper presents the development of a portable, handheld device that optically scans embossed Braille and translates it into printed text in real time. The device addresses a significant communication barrier: while many blind and vision-impaired people read and write using Braille, most sighted people — including teachers of blind students in mainstream classrooms and colleagues in workplaces — cannot read it. The device works by rolling a handheld scanner along a Braille line. A camera system consisting of a linear CCD photodiode array, focusing lens, and an illumination source (four LEDs at right angles to the page) captures shadows cast by the raised Braille dots. A selfoc micro-lens array focuses the dot shadow images onto the CCD sensor. The system runs on a DSP (digital signal processor) that handles motion detection, camera frame sampling, angular misalignment correction, Braille cell compilation, and translation. Each Braille cell is assigned a Binary Coded Braille Cell (BCBC) number representing its dot combination, which is then converted to text using decompression algorithms with switchable lookup tables supporting different Braille codes (mathematics, Greek, English, etc.). Output is displayed on an LCD screen or sent to text-to-speech hardware.
Key findings
The device successfully achieves real-time portable Braille-to-text translation using a novel optical approach. The dot recognition algorithm uses fuzzy logic to estimate whether a Braille dot is present at a given position in the camera frame, making it tolerant of variations in dot positioning and compensating for paper discolouration with age. Reference magnitude samples taken from the edges of the camera field allow the device to work with different types of Braille paper and handle situations where dot shadows affect the reference region. The scanner lens moves above the Braille cell during scanning rather than making contact, causing no wear or tear on the Braille medium. The system compensates for the fact that a handheld scanner cannot maintain perfectly constant speed or alignment — the DSP algorithm handles angular misalignment correction and is tolerant of varying dot locations within camera frames. The Braille cell compilation stage builds complete cells by combining dot presence data from both sides of each cell, and the translation stage supports multiple Braille codes through switchable lookup tables based on Paul Blenkhorn's state machine algorithm.
Relevance
This device addresses a practical and often overlooked accessibility challenge: bidirectional communication between Braille users and sighted non-Braille readers. While much assistive technology focuses on making print accessible to blind people, relatively little attention has been paid to making Braille accessible to sighted people. This is especially relevant in inclusive education settings where blind students produce homework and assessments in Braille that their teachers cannot read, and in workplaces where Braille-literate employees need written communication understood by sighted colleagues. The portable, real-time nature of the device distinguishes it from flatbed scanner-based solutions that lack portability and immediacy. Although the specific hardware has been superseded by advances in camera technology and machine learning-based optical Braille recognition, the communication need the device addresses — bridging the gap between Braille and print literacy — remains relevant and underserved.
Tags: braille · optical braille recognition · assistive technology · visual impairment · communication · portable device