Title of Project: BrailleSync
Group members :-
Roll no. Name
37 Madhura Sanjay Katare
40 Diksha Nitin Khare
56 Antara Anand Sartale
57 Kashish Umar Shaikh
• Abstract :
Visually impaired individuals often rely on Braille for reading and writing, but
communication gaps arise when others cannot understand Braille. To bridge this
gap, we propose a Braille Translator system that converts Braille input into
readable and audible text. The system uses image processing and pretrained
machine learning models to recognize Braille patterns from scanned or captured
images, maps them to corresponding characters through a Braille dictionary, and
provides the output as text and speech. Additionally, it supports reverse
translation by converting normal text into Braille, enabling two-way
communication. With multilingual support and export options, the solution
enhances accessibility, independence, and inclusivity for visually impaired
individuals in education, workplaces, and daily life.
• Rationale:
This project empowers visually impaired individuals by breaking down
communication and information barriers. It acts as a bridge, converting digital
text and speech into a tactile format, thereby providing easy access to a vast
world of information. This not only enhances literacy and educational
opportunities but also promotes greater independence and social inclusion,
allowing individuals to navigate their daily lives with confidence.
• Problem Definition:
Visually impaired individuals face significant barriers in accessing information
and communicating with sighted people because the majority of content is in a
visual format. Braille materials are scarce, expensive, and not easily portable,
creating a fundamental gap in literacy, education, and independence in a
visually-oriented world.
• Introduction:
The Braille Translator is a project designed to tackle the significant accessibility
challenges faced by the visually impaired community. While braille is a
fundamental tool for literacy, the widespread reliance on visual media—from
printed books to digital screens—creates a major barrier. This project leverages
technology to bridge that gap by converting standard text and spoken words into
braille. The core objective is to empower visually impaired individuals with
instant, portable, and affordable access to a vast array of information, from
digital documents and e-books to everyday text. By automating this process, the
project will not only make braille more accessible but also foster greater
inclusion and independence for its users in education, employment, and daily
life.
• Literature Survey:
• Li et al. [1] introduced the DSBI dataset for Braille dot detection, highlighting
challenges in recognizing double-sided embossed documents and the need for
robust algorithms.
• The DSBI dataset [2] provides annotated Braille images and is widely used as a
benchmark for training and testing recognition systems.
• The Liblouis Project [3] developed an open-source Braille translator that
supports multiple languages and Grade-2 Braille, commonly used in accessibility
tools.
• Proposed methodology of solving identification problem:
Input Modules Translation Engine Output Modules
Speech to text module Digital file output
Speech Input
Text normalization Output
OCR and formatting
Text Input formatting
Braille Encoding
Module(Grade 1/2)
Image Input Refreshable Braille
Display output
User interface
Solving the identification problem for a Braille translator project, which involves
accurately recognizing input data like Braille dots, text characters, or spoken words,
follows a systematic methodology. This begins with robust data acquisition and pre-
processing, where diverse inputs are collected, cleaned, and segmented. Next, feature
extraction identifies distinguishing characteristics, which are then fed into model
training and classification using advanced machine learning to accurately map them to
the desired output. Finally, post-processing and error correction refine the
classifications, with continuous evaluation and refinement ensuring the system's
accuracy and robustness.
• Resources and consumable required:
Hardware: PC (i3/4GB RAM min, i5/8GB+), Camera/Scanner, optional GPU.
Software: HTML, CSS, Python 3.x, OpenCV, TensorFlow/PyTorch, NumPy, Pandas,
pyttsx3(TTS),React,Flask.
Consumables: Braille datasets (Kaggle/GitHub)
• Future Enhancement:
Expanded Language & Code Support: Multi-language capabilities and specialized
Braille codes.
Improved Usability: Real-time bi-directional translation, voice control,
personalization.
• Action Plan:
Sr
Week From To Description
no.
Define detailed project scope, Set up development
1 1st 18/8/2025 22/8/2025
environment, Research existing Braille
libraries/algorithms for core translation
Develop the module for converting standard ASCII
2 2nd 25/8/2025 29/8/2025 characters to Grade 1 Braille cell patterns, Create
initial unit tests for character-to-Braille mapping
Begin implementing Grade 2 Braille contraction
3 3rd 1/9/2025 5/9/2025 rules. Address basic capitalization and numeric
indicators.
Continue implementing more Grade 2 contractions
4 4th 8/9/2025 12/9/2025 Integrate Grade 1 and Grade 2 logic into a single
translation core
Develop a module for accepting text input directly
5 5th 15/9/2025 19/9/2025 via a keyboard. Begin designing the initial User
Interface (UI) for text input.
Implement functionality to upload simple text files
6 6th 22/9/2025 26/9/2025 Develop a file parsing module that extracts plain
text from uploaded files.
Develop the module to take the translated Braille
7 7th 29/9/2025 3/10/2025 and format it into a .brf or a plain .txt file.
Implement download functionality.
Develop a basic, functional UI that connects the
8 8th 6/10/2025 10/10/2025 input, translation, and digital file output modules.
Conduct initial end-to-end testing.
Thorough testing of all implemented features.
9 9th 13/10/2025 17/10/2025 Identify and fix bugs, related to contraction rules,
special characters. Error handling, user feedback.
Name and Sign of Group Members Guide Sign
Madhura Sanjay Katare -
Diksha Nitin Khare -
Antara Anand Sartale -
Kashish Umar Shaikh -