DEVELOPMENT OF AN ALPHABETIC CHARACTER RECOGNITION SYSTEM USING MATLAB FOR BANGLADESH

International Journal of Scientific and Research Publications, Volume 9, Issue 1, January 2019 228
ISSN 2250-3153
http://dx.doi.org/10.29322/IJSRP.9.01.2019.p8530 www.ijsrp.or
Development of an Alphabetic Character Recognition
System Using Matlab for Bangladesh
Mohammad Liton Hossain*
, Tafiq Ahmed**, S.Sarkar***
, Md. Al-Amin****
*
Department of ECE, Institute of Science and Technology, National University
**
Electrical Engineering, University of Rostock, Germany
DOI: 10.29322/IJSRP.9.01.2019.p8530
http://dx.doi.org/10.29322/IJSRP.9.01.2019.p8530
Abstract- Character recognition technique, associates a symbolic
identity with the image of the character, is an important area in
pattern recognition and image processing. The principal idea here
is to convert raw images (scanned from document, typed,
pictured etcetera) into editable text like html, doc, txt or other
formats. There is a very limited number of Bangla Character
recognition system, if available they can’t recognize the whole
alphabet set. Motivated by this, this paper demonstrates a
MATLAB based Character Recognition system from printed
Bangla writings. It can also compare the characters of one image
file to another one. Processing steps here involved binarization,
noise removal and segmentation in various levels, features
extraction and recognition.
Index Terms- OCR, Character Recognition, MATLAB, Cross-
Correlation, ImageProcessing.
I. INTRODUCTION
Character Recognition began as a field of research in pattern
identification, artificial neural networks and machine learning.
The different areas covered under this general term are either on-
line or off-line CR, each having its dedicated hardware and
recognition methods. In on-line character recognition
applications, the computer recognizes the symbols as they are
drawn. The typical hardware for data acquisition is the digitizing
tablet, which can be electromagnetic, electrostatic, pressure
sensitive etcetera; a light pen can also be used. As the character
is drawn, the successive positions of the pen are memorized (the
usual sampling frequencies lie between 100Hz and 200Hz) and
are used by the recognition algorithm [1]. Off-line character
recognition is completed afterward the inscription or printing is
performed. Examples of it are usually distinguished: magnetic
and optical character recognition. In magnetic character
recognition (MCR), the fonts are written with magnetic disk and
are designed to modify in a unique way a magnetic field created
by the acquisition device. MCR is mostly used in banking
applications, for instance reading bank checks, because
overwriting or overprinting these characters does not affect the
accuracy of the recognition. While in optical character
recognition, which is the field investigated in this paper, deals
with recognition of characters acquired by optical means,
typically by a scanner or a camera. The symbols can be separated
from each other or belong to structures like words, paragraphs,
figures, etc. They can be printed or handwritten, of any size,
shape, or orientation [2, 3]. Bangla, the mother tongue of
Bangladeshis, is one of the most popular languages in the world,
For the Indian subcontinent Bangla is the Second most popular
language. Moreover 200 million people of eastern India and
Bangladesh use this language and also the institutional language
in our country [1]. So recognition of Bangla alphabet is always a
special interest to us considering the massive number of official
papers are being scanned and processed every day. This work
aims to develop a system that categorizes a given input (Bangla
Characters) as belonging to a certain class rather than to identify
them uniquely, as every input pattern. It also performs character
recognition by quantification of the character into a mathematical
vector entity using the geometrical properties of the character
image. Global usage of this system may reduce the hard labor of
the concerning government employees every day.
II. BACKGROUND
Basic Bangla character set comprises 11 vowels, 39 consonant
and 10 numerals. There are also compound characters being
combination of consonant with consonant as well as consonant
with vowel [1, 4]. The complete set of characters, need to
recognize by the proposed recognition system, is provided in
Figure1.
Figure1: The whole character set of Bangla Alphabet
A. OVERVIEW OF THE IMAGE TYPES AND
PROPERTIES
The RGB (true-color) images are put in storage as 3 distinct
image matrices; one of red (R) in each pixel, one of green (G)
and one of blue (B); similar to the wavelength-sensitive cone
cells of human eye. While in a grayscale or graylevel digital
image, the value of each pixel is a single value, carries only
intensity information. A grayscale image is not the black and
white Binary (2-leveled) image. Gray image is represented by
black and white shades or combination of levels. 8 bit gray image

ISSN 2250-3153
represents total 2^8 levels, from black to white, 0 equals black
and 255 is White. While Binary images, have only two probable
values for individual pixel, are produced by thresholding a
grayscale or color image to separate an object in the image from
its background [5].
Also, color images are generally described by using the
following three terms: Hue, Saturation and Lightness. Hue is
another word for color and dependent on the wavelength of light
being reflected or produced. Saturation refers to that how pure
or intense a given hue is. 100% saturation means there’s no
addition of gray to the hue - the color is completely pure [15].
Lightness measures the relative degree of black or white that’s
been mixed with a given hue. Adding white makes the color
lighter (creates tints) and adding black makes it darker (creates
shades). The effect of lightness is relative to other values in the
composition [6, 15].
III. METHODOLOGY
The proposed Bangla Chracter Recognition system is
developed here by the following steps:
Step1: Printed Bangla Script input to Scanner
Step2: Raw scanned Document from Scanner
Step3: Convert the scanned RGB image into Grayscale
image
Step3: Convert Grayscale image to Binary image.
Step4: Segmentation and Feature Extraction.
Step5: Classification and Post-processing.
Step6: Output editable text.
A scanned RGB color image here is input to the system. The
proposed text detection, recognition and translation algorithm
(shown in Figure2) consists of following steps: Binarization,
Noise removal, Segmentation, Features extraction, Correlation
and Cross-Correlation [1, 7].
A. Pre-Processing
Binarization
Binarization is the technique here by which the gray scale
pictures are transformed to binary images to facilitate noise
removal. Typically the two colors used for a binary image are
black and white though any two colors can be used. The color
used for the object in the image is the foreground color while the
rest of the image is the background color . Binarization separates
the foreground (text) from the background information [1-2].
Figure 2: An overview of the proposed Bangla character recognition
system.
Noise Removal
Scanned documents often contain noise that arises due to the
accessories of printer or scanner, print quality, age of the
document, etc. Therefore, it is necessary to filter this noise before
processing the image. This low-pass filter should avoid as much
of the distortion as possible while holding the entire signal [1, 8].
B. Segmentation
Image segmentation is the process of partitioning a digital image
into multiple segments (sets of pixels, also known as super
pixels). The goal of segmentation is to simplify and/or change
the representation of an image into something that is more
meaningful and easier to analyze. Image segmentation is the
process of assigning a label to every pixel in an image such that
pixels with the same label share certain characteristics or
property such as color, concentration, or quality. The result of
image segmentation is a set of sections that collectively cover the
whole image, or a set of lines take out from the image (or edge
detection). Segmentation of binary image is achieved in altered
levels contains line segmentation, word segmentation, character
segmentation. Thresholding often delivers an easy and suitable
method to accomplish this segmentation on the base of the
different concentrations or colors in the foreground and
background areas. In a single pass, each pixel in the image is
related with this threshold value. If the pixel's intensity is higher
than the threshold, the pixel is set to white in the output,
corresponding foreground. If it is less than the threshold, it is set
to black, indicating background [1, 2, 8].
Features Extraction
Feature taking out is the special form of dimensional decrease
that proficiently characterizes exciting parts of an image as a
solid feature route. This method is suitable when image sizes are
huge and a condensed feature illustration is essential to swiftly
complete tasks, such as image matching and recovery. Feature
detection, feature mining, and matching are often combined to
solve common computer vision problems such as object
detection and recognition, content-based image retrieval, face
detection and recognition, and texture classification [1, 9].

ISSN 2250-3153
C. Correlation and Cross-Corellation
Image correlation and tracking is an optical technique that works
tracking and image registration method for accurate 2D and 3D
dimension of changes in images. Cross-correlation is the
measure of similarity of two images. Each of the images is
divided into rectangular blocks. Each block in the first image is
correlated with its corresponding block in the second image to
produce the cross-correlation value as a function of position
[10,11].
IV. MATLAB IMPLEMENTATION
An implementation is the realization of a technical specification
or algorithm as a program, software component, or other
computer system through computer programming and
deployment. The proposed system is implemented in MATLAB
and the accuracy is also tested. Using MATLAB ‘imread’
function of, a noisy scanned RGB image is loaded to the input
system (Figure.3). After performing Grayscale conversion by
‘rgb2gray’, Figure.4 is thus obtained; ‘rgb2gray’ eliminates the
hue and saturation information of the RGB image while retaining
the luminance. Figure.5-6 shows outcome after binarization
(using ‘im2bw’) and noise-removal from this black and white
image. Figure.7-8 shows the process of character segmentation
[12-13].
Performing cross-correlation between input image and template
image, the peak point is found (Figure.9). Peak point is the point
that indicates the highest matching area based on cross-
correlation. Thus, the appropriate character is identified
(Figure.10). Here Bangla Unicode Characters’ hex value is used
to make the recognized character machine-readable. Recognized
character is here displayed in MATLAB built in browser [14]. Figure 9: Marked Peak in cross correlation refer Character Detection
Figure 10: Recognized character from Template.
Figure 7: Image information for segmentation Figure 8: Character after Segmentation.
0 represented by Black & 1 represents White.

ISSN 2250-3153
V. GUI (GRAPHICAL USER INTERFACE)
DEVELOPMENT
A user-friendly Graphical User Interface is also designed for this
system, where user can insert an image to recognize the desired
character. User can input any image using “Load Image” option.
After loading an image, user can select a character using mouse
pointer, crop image and continue processing using “Image
Processing” option. Following the algorithms, the system would
recognize the selected character and provide it in editable format
if “Recognize Character” command is executed from the options.
VI. ANALYSIS
The recognition rate of the proposed recognition method is
remarkable with the accuracy rate of almost 98-100%. Image
processing toolbox and built in MATLAB functions have been
quite helpful here for detection of single Bangla character and
digit. The input images can be taken from any optical scanner or
camera. The output allows saving recognized character as html,
txt etcetera format for further processing. The only limitation of
the system is the fixed font size of the character. Further
improvement for all possible font type and size is reserved for
the future. Also it is developed only for single Bangla character
recognition at a time. Multiple characters recognition along with
connected letters is also left open for the future. Connecting the
program algorithm with the neural network may also be helpful.
After the successful review and payment, IJSRP will publish
your paper for the current edition. You can find the payment
details at: http://ijsrp.org/online-publication-charge.html.
VII. CONCLUSION
Complex character based language like Bangla clearly demands
deep research to meet its goals. Recognition of Bangla characters
is still in the intermediate stage. The proposed system is
developed here using template matching approach to recognize
individual character image. Even though the user-friendly GUI
gives several advantages to individual users, this system is still
facing a number of limitations which is quite considerable.
Improving the system to suit handwritten characters, all printed
fonts in considerable amount of time with high accuracy is the
recommended future work.
ACKNOWLEDGMENT
At first all thanks goes to almighty creator who gives me the
opportunity, patients and energy to complete this study. I would
like to give thanks to Mr. Masudul Haider Imtiaz; Assistant
Professor at Institute of Science and Technology. I have found
always immense support from him to keep my work on the right
way. His door was always open for me, when I have got myself
in trouble.
Finally, I must express my very profound gratefulness to my
Figure 11: Computer editable Character
Figure 12: A specific instance of character recognition using
developed MATLAB GUI

ISSN 2250-3153
parents and to my wife for providing me with constant support
and encouragement during my years of study and through the
process of researching and writing this paper. This
accomplishment would not have been possible without them.
Thank you.
REFERENCES
REFERENCES
[1] Md. Mahbub Alam and Dr. M. Abul Kashem, A Complete Bangla
OCR System for Printed Characters, JCIT, 01(01).
[2] MCR, OCR, Binary Image, Image Segmentation, Feature-extraction,
Wikipedia, the free encyclopedia.
[3] Sandeep Tiwari, Shivangi Mishra, Priyank Bhatia, Praveen Km.
Yadav, Optical Character Recognition using MATLAB, International
Journal of Advanced Research in Electronics and Communication
Engineering (IJARECE) ,5(2), (2013).
[4] Bengali Alphabet, Wikipedia, the free encyclopedia.
[5] RGB Color Model, Grayscale, Wikipedia, the free encyclopedia.
[6] Hue, saturation and brightness, Wikipedia, the free encyclopedia.
[7] Md. Alamgir Badsha, Md. Akkas Ali, Dr. Kaushik Deb, Md.
Nuruzzaman Bhuiyan, Handwritten Bangla Character Recognition
Using Neural Network, International Journal of Advanced Research in
Computer Science and Software Engineering, 11(2), (2012).
[8] Samiur Rahman Arif, Bangla Character Recognition using Feature
extraction, Project work, Department of Computer Science, Brac
University, (2007).
[9] Ahmed Asif Chowdhury et.al Optical Character Recognition of
Bangla Characters using neural network: A better approach,
Department of Computer Science, BUET, Bangladesh.
[10] Amin Ahsan Ali & Syed Monowar Hossain, Optical Bangla Digit
Recognition using Back propagation Neural Networks, Project work,
Department of computer science, University of Dhaka, (2003).
[11] Definition, Cross-correlation, Online.
[12] MATLAB, Wikipedia, the free encyclopedia.
[13] MATLAB, Image processing toolbox, Mathworks, Online.
[14] Rafael C. Gonzalez,Richard E. Woods, Setven L. Eddins, Digital
Image Processing using MATLAB, 2nd Edition, (2009).
[15] The Fundamentals of Color: Hue, Saturation, And Lightness,
available at online: https://vanseodesign.com/web-design/hue-
saturation-and-lightness/
AUTHORS
First Author –– Mohammad Liton Hossain, Lecturer,
Department of ECE, IST
litu702@gmail.com
+8801768346307
Second Author -Tafiq Ahmed, Electrical Engineering, University of
Rostock, Germany
Third Author –S.Sarkar, Department of ECE, IST

DEVELOPMENT OF AN ALPHABETIC CHARACTER RECOGNITION SYSTEM USING MATLAB FOR BANGLADESH

Recommended

More Related Content

What's hot (17)

Similar to DEVELOPMENT OF AN ALPHABETIC CHARACTER RECOGNITION SYSTEM USING MATLAB FOR BANGLADESH (20)

More from Mohammad Liton Hossain (9)

Recently uploaded (20)

DEVELOPMENT OF AN ALPHABETIC CHARACTER RECOGNITION SYSTEM USING MATLAB FOR BANGLADESH