Projection Profiling Based Sinhala Braille Character Recognition and Conversion
View/ Open
Date
2020Author
Ariyarathna, W Randika
Kahandagamage, L Rasika
Kumara, WMP Wasantha
Metadata
Show full item recordAbstract
Abstract: The instantaneous conversion of
written Braille characters to readable,
colloquial text of the particular language it
was written in, has been a hurdle for many
sighted personal. Various solutions have
been proposed in this regard and many
solutions are available to convert Braille to
different languages. However, there are very
few attempts to do the same for Sinhala
Braille symbols and at present widely using
conversion method is to use an individual
well versed in both Sinhala language as well
as Sinhala Braille characters. As such, a
considerable delay occurs in the translation
process. Hence the visually impaired
academics and students, as well as soldiers
visually impaired due to war, who work with
Braille encounter many hardships. Their
teachers and superiors also have faced
numerous problems as they are unable to
understand Braille. One such consequence is
that students sitting for exams face delays in
receiving exam results as translation from
Braille to Sinhala introduces additional
delays in addition to often the context being
lost in translation.
Proposed conversion engine will easily
convert the Sinhala Braille text on papers to
the corresponding Sinhala text and will assist
the learning and communication processes of
visually impaired citizens of Sri Lanka. The
Braille printed documents to be converted
through this software need to be subjected to
image pre-processing, projection profiling
and Braille recognition processes, capturing
a picture of the Braille characters on a sheet
initiates the process. Six dots are used to
create Braille symbols. This system identifies
these distinct symbols using a specially
constructed framework and algorithms,
followed by comparing them against a
database of symbols in the recognition
process and the final output is derived using
the Sinhala alphabet.
Collections
- Computer Science [66]