Professional Documents
Culture Documents
Project Report
TEAM MEMBERS
19BCE2659: RAVI KASAUDHAN
19BCE2664: AAYUSH PARAJULI
S.No Title of the Year Journal name Methodology Performanc Data set / Evaluation
resource / Conference (Key algorithms / e metrics Data used in observations/
(journal title/ approach) the resource
paper/conference Website link Comments
paper/ title of the
web page)
1 Development of 2021 International The paper System has There are 64 This
a Text to Braille Journal for represents performed possible application
Interpreter Research in complete extraction of combination converts an
through Optical Engineering methodology on text from an s as there are image
Image Processing Application how scanned input image no dots for document to
& images and how it and then word space. Braille
Management is translated into passing the Braille is a format and
(IJREAM) braille format. The text for series of thus this
scanned images conversion characters or application
are pre-processed into Braille "cells" can help a lot
and converted to Language. consisting of of visually
grayscale and then This process six extended impaired
passed through an benefits to point people. This
adaptive threshold schools, patterns and System will
function for parents or arranged in be an Open
conversion to anyone who a rectangle Source. So
binary image. can help with two that anyone
Translation to blind columns of can have
Braille includes people. three points. access to it as
conversion of there will be
numbers, no need of
alphabets, subscriptions
symbols and
compound letters, or purchase
translated text can required.
then be saved for
printing.
2 IEEE The camera will Percentage There were Despite the
DEVELOPMEN 2018 capture an image, accuracy 180 tests existence of
T OF A TEXT in tiff format, and Average gathered in different
TO BRAILLE which will be accuracy this study, laws,
INTERPRETER analyzed through was used as 30 tests each organizations
FOR PRINTED an optical image performanc for the fonts , and
DOCUMENTS processing e metrics. Times New movements
THROUGH integrated in a Roman, to help the
OPTICAL GUI. After Arial, and visually-
IMAGE analysis, a Futura for impaired
PROCESSING universal word-per- people to
asynchronous word mode cope up with
receiver/transmitte and this time and
r (UART) device continuous age, there is
will interface the mode. still a
communication Checking hindrance for
between the was done visually-
computer and the manually by impaired
microcontroller. letting the people to
The respondents have a fully
microcontroller spell out accessible
then sends the what they education
data to the read. due to
refreshable braille limitations of
cells. resources.
3 Text Translation Indian The translation of The A separate This paper
of Scanned Hindi 2017 Journal of scanned Hindi text performanc database of successfully
Document to Science and into Braille code e metrics Hindi presented the
Braille via Image Technology involves the were consonants conversion of
Processing following four analyzed on and matras scanned
processes: (i) the basis of is generated. Hindi text
Hindi database accuracy of A scanned into the
generation of image correspondin
letter
consonants and containing g Braille
matching
matras using all the Hindi using PCA
image with hindi alphabets and image
segmentation. database along with segmentation
(ii) Segmentation and hindi to matras is technique.
of test images into braille segmented The first part
lines followed by conversion. to obtain of the paper
words and finally individual is a
into letters. letters. generation of
(iii) Letter These Hindi
matching of individual database
segmented letters Hindi letters where the
with the generated of size 187 x size of every
Hindi database 128 are image is
using Principal stored in 187x128
Component MATLAB pixels which
Analysis (PCA). to generate a require very
(iv) Conversion of Hindi less memory
matched letter into database. of the
corresponding system.
Braille code.
4 CONVERSION IEEE This paper The model Custom The Support
OF HINDI 2020 explores the was run for dataset of Vector
BRAILLE TO conversion of several 34800 machine
SPEECH USING Devanagari Hindi times to find images was (SVM)
IMAGE AND Braille, first to that it would created by created had
SPEECH text, and be suitable taking the highest
PROCESSING subsequently to at 15 pictures of accuracy.
speech. The epochs.The braille texts
authors achieved loss and from a book
an accuracy of accuracy for using a
100 percent using random and phone
the conventional stratified camera.The
method. Also splits were dataset
validated the measured consists of
quality of Hindi and a macro 600 images
speech generated average for each of
from the text-to- performance the 58
speech model. metrics braille
table was characters.
used for
both the
splits.
Finally the
percentage
of accuracy
was checked
for all the
SVM
approach,
Alex
net(random
split and
stratified
split)
V. PROPOSED RESEARCH FRAMEWORK
a. RESEARCH ARCHITECTURE:
Our project is principally divided into two sections. the primary section is
especially concerned with preprocessing of a picture and image to string
conversion. The second section deals with translating the resultant output
text from image to Grade 2 braille. The important thing that ought to be
taken into consideration is to get a picture suitable for extraction of strings
from them. For this we have applied many pre-processing image
techniques which are rescaling the input image, converting the image into
grayscale, eliminating any shadows that are present on the image. For the
noise removal dilation and erosion is implemented. Finally, image
thresholding is executed to maintain the intensity of black and white
pixels. This follows to section two where nation characters are recognized
and mapped to the corresponding braille characters using tesseract.
Fig: Research Architecture
2. Binarization
In our project we've used Otsu’s algorithm for binarization. Otsu’s
thresholding method corresponds to the linear discriminant criteria that
assumes that the image consists of only object (foreground) and
background, and therefore the heterogeneity and variety of the
background is ignored. Otsu set the brink so on attempt to minimize the
overlapping of the category distributions. This method gives a threshold
for the entire image considering the varied characteristics of the full image
(like lighting conditions, contrast, sharpness etc.) and that threshold is
employed for Binarizing image.
3. Noise Removal
a. Erosion
Erosion is one amongst two fundamental operations in morphological
image processing from which all other morphological operations are
based. The erosion operation usually uses a structuring element for
probing and reducing the shapes contained in the input image.
For sets A and B in Z2, erosion is denoted by:
b. Dilation
▪ Handling capitals:
VII. IMPLEMENTATION
PROJECT CODES:
IMAGE TO TEXT
Fig: Input image
Fig: Resultant Text from the input image
Main.py
mapAlphaToBraille.py
mapBrailleToAlpha.py
brailleToAlpha.py
alphaToBraille.py
Printer.py
VIII. RESULT (SCREENSHOTS)
Fig: The text file obtained from the image
IX. CONCLUSION
[1] Padmavathi, S., Reddy, S. S., & Meenakshy, D. (2013). Conversion of braille
to text in english, hindi and tamil languages. arXiv preprint arXiv:1307.2997.
[2] Cruz, J. L. D., Ebreo, J. A. D., Inovejas, R. A. J. P., Medrano, A. R. C., &
Bandala, A. A. (2017, December). Development of a text to braille interpreter for
printed documents through optical image processing. In 2017IEEE 9th International
Conference on Humanoid, Nanotechnology, Information Technology,
Communication and Control, Environment and Management (HNICEM) (pp. 1-6).
IEEE.
[3] Parvathi, K., Samal, B. M., & Das, J. K. (2015, February). Odia Braille: Text
transcription via image processing. In 2015 International Conference on Futuristic
Trends on Computational Analysis and Knowledge Management (ABLAZE) (pp.
138-143). IEEE.
[4] Kaur, P., Ramu, S., Panchakshari, S., & Krupa, N. (2020, November).
Conversion of Hindi Braille to Speech using Image and Speech Processing. In 2020
IEEE 7th Uttar Pradesh Section International Conference on Electrical, Electronics
and Computer Engineering (UPCON) (pp. 1-6). IEEE.
[5] Murthy, V. V., & Hanumanthappa, M. (2018, February). Improving Optical
Braille Recognition in Pre- processing Stage. In 2018 International Conference on
Soft-computing and Network Security (ICSNS) (pp. 1-3). IEEE.
[6] Beg, U., Parvathi, K., & Jha, V. (2017). Text translation of scanned Hindi
document to braille via image processing. Indian journal of science and technology,
10(33).
[7] Pilu, M., & Pollard, S. (2002, September). A light-weight text image
processing method for handheld embedded cameras. In bmvc (Vol. 2, pp. 547-556).
[8] Zhang, S., & Yoshino, K. (2007, September). A braille recognition system by
the mobile phone with embedded camera. In Second International Conference on
Innovative Computing, Information and Control (ICICIC 2007) (pp. 223-223).
IEEE.
[9] Singh, M., & Bhatia, P. (2010). Automated conversion of English and Hindi
text to Braille representation. International Journal of Computer Applications, 4(6),
25-29.
[10] S, S., S K, S., Ranjini, S., N, V., & M C, P. (2018, June). Text Image to Braille
Code Converter. International Journal of Engineering Research in Electronics and
Communication Engineering (IJERECE),5(6),66-69.