Design and implementation of optical character recognition using template matching for multi fonts size

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
_______________________________________________________________________________________
Volume: 04 Issue: 02 | Feb-2015, Available @ http://www.ijret.org 398
DESIGN AND IMPLEMENTATION OF OPTICAL CHARACTER
RECOGNITION USING TEMPLATE MATCHING FOR MULTI FONTS /
SIZE
Nikhil Rajiv Pai1
, Vijaykumar S. Kolkure2
1
M.E. (Electronics, Appeared), Department Of Electronics Engineering,Bharatratna Indira Gandhi College of Engineering,
Affiliated to Solapur University, Solapur, Maharashtra, India.
2
Assistant Professor, Department Of Electronics Engineering, Bharatratna Indira Gandhi College of Engineering, Affiliated to
Solapur University, Solapur, Maharashtra, India.
Abstract
Optical character recognition (OCR) is an efficient way of converting scanned image into machine code which can further edit.
There are variety of methods have been implemented in the field of character recognition. This paper proposes Optical character
recognition by using Template Matching. The templates formed, having variety of fonts and size .In this proposed system, Image
pre-processing, Feature extraction and classification algorithms have been implemented so as to build an excellent character
recognition technique for different scripts .Result of this approach is also discussed in this paper. This system is implemented in
Matlab.
Keywords- OCR, Feature Extraction, Classification
---------------------------------------------------------------------***---------------------------------------------------------------------
1. INTRODUCTION
Basically Optical character recognition (OCR) having two
main types: 1.Off-line Character Recognition 2.On-line
Character Recognition. On-line character recognition is
simple and effective way for recognition, as the input to this
system is data which is collected online. These systems
recognize a character while the user is writing with an online
writing device. Off-line character recognition simply known
as OCR. Off-line character recognition is somewhat difficult
task, as this system has to recognize handwritten characters.
The handwritten characters having different size and fonts. It
varies from person to person. Therefore it is somewhat
complicated to identify a character. This system proposes an
effective approach ‘Template Matching’. In this variety of
templates have been formed with variety of fonts and size.
This method can recognize a character irrespective of its
style, size and font. The input scanned image is compared
with standard templates. The most closely matched template
is declared as matched template. Thus character can
recognized. The input scanned image undergoes the process
of pre-processing including binarization, segmentation, image
resizing etc.
1.1 Binarization
In this stage, the input scanned color image is converted into
grayscale image. The color image is having parameters such
as, R-G-B (Red-Green- Blue). Firstly the R-G-B values of
input scanned image are calculated. After calculation of R-G-
B values, 30% of Red, 59% of Green and 11 % of blue values
are separated. The separated values are added together .The
obtained addition is the expected grayscale value. Then the
grayscale histogram of input scanned image F(x, y) is
calculated. After binarization, the binarized image is
forwarded for segmentation, so as to separate the sub parts of
an input image.
1.2 Image Resizing
After the process of binarization, segmentation etc., the
image should be rearranged in all directions. The image
resizing is done from left, right, top, bottom side of input
image. Once the resizing is done, then templates of different
fonts, styles, size are formed.
1.3 Template Formation and Matching
This is the base of this approach. In this, different templates
are designed in such a way that it is having variety of font
size and styles. Due to this variety, characters can recognize
by the system irrespective of its font size and font style.
Whatever the input characters to be recognized are compared
with these standard templates. First of all, area of template is
calculated from all directions. It includes calculation of
distance between border and boundary of characters. This
calculation of distance between border and boundary of
characters is done for both templates and characters to be
recognized. Whenever the character is to be recognized, a
respective character is marked from all directions that means
from top, bottom, right, left directions. This marking
specifies the boundaries of that respective character. Then
that character is compared with different templates. This
comparison is done for the distance between border and
boundary of characters, which has been already calculated.
Suppose, the distance between border and boundary of a test
character is Td and the same for reference template is Rd.

_______________________________________________________________________________________
Then during comparison, every time the difference between
each Rd and Td is calculated. The minimum difference
between Td and Rd indicates likely matching elements. Thus
the template having minimum distance is considered as best
matched template. In this way a particular character gets
recognized.
2. RESULTS AND DISCUSSION
Here Optical Character Recognition technique uses Template
Matching approach. This system is implemented in Matlab.
Here we have considered 5 different cases. In this, each
character set having different font size, styles etc. The
purpose of considering such different character sets is to test
the accuracy of OCR.
n i k h i l k a v i t a r a j i v p a i
Case 1: constant font size and font style
N i k h i l K A V i t a R a J I V p a i
Case 2: variable font style and constant font size
n i k h i l k a v i t a r a j i v p a i
Case 3: variable font size and constant font style
N i k h i l K A V i t a R a J I V p a i
Case 4: Variable font style and font size
N I K H I L K A V I T A R A J I V P A I
Case 5: Italic, Bold font style and constant font size
The result of above mentioned cases are shown below.
Fig 1: Results of Experiment
Table 1: Recognition rate of different cases
Sr. No. Case Recognition Rate
01 Case 1 95%
02 Case 2 96%
03 Case 3 96%
04 Case 4 93%
05 Case 5 95%
Here we have considered different cases for recognition. In
all cases, the characters are differs in font style and size. As
multi fonts/ size templates have been designed in this
approach, the above characters get compared with the
standard templates. By comparing with the templates, it has
been easy task to recognize a character for the system.
Therefore in all cases accuracy of OCR is above 90%.
3. CONCLUSION
For different cases of characters, OCR system has been
implemented. As this proposed system is designed with the
standard templates, characters can recognize effectively and
more efficiently by comparing them with templates. OCR
using template matching can identify the characters more
accurately irrespective of its font size and style. For
mentioned cases OCR gets recognition rate above 90%. Thus
Optical Character Recognition using Template Matching can
become a smooth way for character recognition.
0
0.2
0.4
0.6
0.8
1
1.2
Case case 1 case 2 case 3 case 4 case 5
SR. No. 1 2 3 4 5
Series1
Series2

_______________________________________________________________________________________
REFERENCES
[1]. Jagruti Chandarana, Mayank Kapadia Optical Character
Recognition International Journal of Emerging Technology
and Advanced Engineering Volume 4, Issue 5, May 2014
[2]. Binod Kumar Prasad, GoutamSanyal A model Approach
to Off-line English Character Recognition International
Journal of Scientific and Research Publications, Volume 2,
Issue 6, June 2012
[3]. SandeepTiwari , Shivangi Mishra, Priyank Bhatia,
Praveen Km. Yadav[May 2013] Optical Character
Recognition using MATLAB International Journal of
Advanced Research in Electronics and Communication
Engineering (IJARECE)Volume 2, Issue 5, May 2013
[4]. Majida Ali Abed Hamid Ali Abed Alasadi Simplifying
Handwritten Characters Recognition Using a Particle Swarm
Optimization Approach European Academic Research,
Volume I, Issue 5/ August 2013
[5]. Mahesh Goyani, Harsh Dani, Chahna Dixit Handwritten
Character Recognition – A Comprehensive Review
International Journal of Research in Computer and
Communication Technology, Volume 2, Issue 9, September -
2013
[6]. Sushree Sangita Patnaik and Anup Kumar Panda Particle
Swarm Optimization and Bacterial Foraging
OptimizationTechniques for Optimal Current Harmonic
Mitigation by Employing Active Power Filter Applied
Computational Intelligence and Soft Computing Volume
2012, Article ID 897127.
[7]. Pritpal Singh, Sumit Budhiraja Feature Extraction and
Classification Techniques in O.C.R. Systems for Handwritten
Gurumukhi Script – A Survey International Journal of
Engineering Research and Applications (IJERA) Volume 1,
Issue 4, pp. 1736-1739.
[8]. Lipi Shah, Ripal Patel, Shreyal Patel, Jay Maniar Skew
Detection and Correction for Gujarati Printed and
Handwritten Character using Linear Regression International
Journal of Advanced Research in Computer Science and
Software Engineering
Volume 4, Issue 1, January 2014
[9]. Ritesh Kapoor, Sonia Gupta, and C.M. Sharma, “Multi-
font/size character recognition and document scanning,” Int.
J. of Computer Application, volume 23, no.1, pp. 21-24,
2011.
BIOGRAPHIES
Nikhil Rajiv Pai 1
currently pursuing M.E.
(Electronics) From Bharatratna Indira
Gandhi College of Engineering, Solapur,
Maharashtra, India. His area of interest is
image processing, MATlab.
Vijaykumar S. Kolkure 2
has completed M.E. (Electronics)
from W.C. Sangli, Maharashtra, India. He has 10 years of
teaching experience. Currently he is working as Assistant
Professor at Bharatratna Indira Gandhi College of
Engineering, Solapur, Maharashtra, India. His area of interest
is Image Processing, Video Processing.

Design and implementation of optical character recognition using template matching for multi fonts size

Recommended

More Related Content

What's hot (20)

Viewers also liked (19)

Similar to Design and implementation of optical character recognition using template matching for multi fonts size (20)

More from eSAT Journals (20)

Recently uploaded (20)

Design and implementation of optical character recognition using template matching for multi fonts size