International Journal ofof Computer Science and Engineering Research Development (IJCSERD), ISSN 2248International Journal Computer Science

and Engineering Research and 9363Development2248-9371 (Online) Volume 1, Number 2, ISSN 2248- (2011) and (Print), ISSN (IJCSERD), ISSN 2248-9363 (Print), May-October

IJCSERD

9371 (Online), Volume 1, Number 2, May-October (2011) pp. 75-83 PRJ Publication http://www.prjpublication.com/IJCSERD.asp

PRJ PUBLICATION

HANDWRITTEN CHARACTER RECOGNITION USING MULTIRESOLUTION TECHNIQUE AND EUCLIDEAN DISTANCE METRIC
1

D.K. Patel1, T. Som1, S. K. Yadav2 Department of Applied Mathematics, Institute of Technology, Banaras Hindu University, Varanasi - 221005, India 2 Department of Statistics, Faculty of Science, Banaras Hindu University, Varanasi - 221005, India E-mail: 1dkpatel.bhu@gmail.com, 1tsom.apm@itbhu.ac.in, 2sush.stats@gmail.com

ABSTRACT In the present paper, the widely common problem of handwritten character recognition has been tackled with multiresolution technique using discrete wavelet transform and Euclidean distance metric. The technique has been tested and found to be more accurate and economic in respect of the recognition process time of the system. Characters are classified into 26 pattern classes based on appropriate properties. Features of the handwritten character images are extracted by discrete wavelet transform used with appropriate level of multiresolution technique, then each pattern class is characterized by a mean vector. Distances from input pattern vector to all the mean vectors are computed by Euclidean distance metric. Minimum distance determines the class membership of input pattern vector. The proposed method provides good recognition accuracy of 90% for handwritten characters even with less no. of samples. Keywords: Discrete Wavelet Transform, Euclidean Distance Metric, Feature Extraction, Handwritten Character Recognition, Bounding Box, Mean Vector.
I. INTRODUCTION

Handwritten Character Recognition (HCR) is an area of pattern recognition process that has been the subject of considerable research during the last few decades. The ultimate objective of any HCR system is to simulate the human reading capabilities so that the computer can read, understand, edit and do similar activities as human do with the text. Mostly, English language is used all over the world for the communication purpose, also in many Indian offices such as railways, passport, income tax, sales tax, defense and public sector undertakings such as bank, insurance, court, economic centers, and educational institutions etc. A lot of works of handwritten English character recognition have been published but still minimum training time and high recognition accuracy of handwritten English character recognition is an open problem. Therefore, it is of great importance to develop automatic handwritten character recognition system for English language. In this paper, efforts have been made to develop automatic
75

International Journal of Computer Science and Engineering Research and Development (IJCSERD), ISSN 22489363 (Print), ISSN 2248-9371 (Online) Volume 1, Number 2, May-October (2011)

handwritten character recognition system for English language with high recognition accuracy and minimum classification time. Handwritten character recognition is a challenging problem in pattern recognition area. The difficulty is mainly caused by the large variations of individual writing style. To get high recognition accuracy and minimum classification time for handwritten character recognition, we have applied multiresolution technique using discrete wavelet transform and Euclidean distance metric. Experimental results show that the proposed method used in this paper for handwritten English character recognition is giving high recognition accuracy and minimum classification time. In what follows we briefly describe the different techniques used in our paper. A. Multiresolution Representation of images in various degrees of resolution is known as multi resolution process. In multiresolution process, images are subdivided successively into smaller regions [1-2]. Wavelets are used as the foundation of multi resolution process. In 1987, wavelets are first shown to be the foundation of a powerful new approach to signal processing and analysis called multiresolution theory [2]. Multiresolution theory is concerned with the representation and analysis of images at more than one resolution. We use this technique in HCR system. B. Euclidean Distance Metric The Euclidean distance metric is significantly simplified under the following assumptions [3-4]: The classes are equiprobable. The data in all classes follow Gaussian distributions. The covariance matrix is the same for all classes. The covariance matrix is diagonal and all elements across the diagonal are equal.

Given an unknown pattern vector x , assign it to class w j if

x mi

( x m i ) T ( x mi ) x m j

, i j

Euclidean distance metric is often used, even if we know that previously stated assumptions are not valid, because of its simplicity. It assigns a pattern to the class whose mean is closed to it with respect to the Euclidean norm [5].
II. METHODOLOGY OF THE PROPOSED HANDWRITTEN CHARACTER RECOGNITION SYSTEM

This paper is divided into four sections. In section A, samples of handwritten characters are collected. In section B, some preprocessing steps are performed on the collected samples of handwritten characters. In section C, pattern vectors are generated using multiresolution
76

International Journal of Computer Science and Engineering Research and Development (IJCSERD), ISSN 22489363 (Print), ISSN 2248-9371 (Online) Volume 1, Number 2, May-October (2011)

technique by applying discrete wavelet transform on preprocessed handwritten characters. In section D, mean vectors are computed of each class and then Euclidean distance metric is used to compute distances from input pattern vector to all the mean vectors and input handwritten character is recognized by the minimum distance obtained in section D. A complete flowchart of handwritten English character recognition system is given in the following figure [6].

Scanned Handwritten Character

Character Resized into 100 by 100 Pixels

Binarization, Skeletonization, and Edge Detection

Preprocessing

Slant Angle Removal of Character Image

Character Image after Slant Correction

Multiresolution Using Wavelet Transform

Feature Extraction

Preparation of Pattern Classes and Mean Vector

Pattern Classes and Mean Vector

Minimum Distance Classifier

Euclidean Distance Metric

Recognition of Input handwritten character

Output

Figure 1 A. Data Collection

A flowchart of handwritten character recognition system

First of all, the data of English characters is collected in the written form on blank papers by people of different age groups. These characters are written by different blue/black ball point pen. The collected samples of handwritten characters are scanned by scanner into JPEG format on 600 ppi. Then all the characters are separated and resized into 100 by 100 pixel images. The example of the samples of handwritten characters is given below.

77

International Journal of Computer Science and Engineering Research and Development (IJCSERD), ISSN 22489363 (Print), ISSN 2248-9371 (Online) Volume 1, Number 2, May-October (2011)

Figure 2

Sample of characters written by 10 different persons

100100 Figure 3

100100

100100

100100

100100

Some separated characters from the sample in Fig. 1 and resized into 100 by 100 pixel images

B. Preprocessing
78

International Journal of Computer Science and Engineering Research and Development (IJCSERD), ISSN 22489363 (Print), ISSN 2248-9371 (Online) Volume 1, Number 2, May-October (2011)

The separated RGB character images of 100 by 100 pixel resolution obtained in the section A are converted into grayscale images and then the grayscale images is again converted into binary images using appropriate gray scale thresholding. Now, binary image is thinned using skeletonization infinite times. Edges of these thinned images are detected using appropriate thresholding and then further dilated using appropriate structure element. These steps are known as preprocessing and preprocessing steps are given in the following figure. Handwritten Character (Original Image Resized into 100 by 100 pixel)

Grayscale Image

Binary Image after Binarization

Dilated Image After Slant Correction Figure 4

Dilated Image

Edge Detection of Thinned Image

Thinned Image

Results of preprocessing steps applying on a handwritten character

C. Feature Extraction In this section, multiresolution technique is applied on the dilated images obtained in section B and this is achieved by applying discrete wavelet transform. In this section, pattern vectors are generated [7-13]. Discrete wavelet transform has given good results in different image processing applications. It has excellent spatial localization and good frequency localization properties that makes it an efficient tool for image analysis. There are different multiresolution techniques such as image pyramids, subband coding and discrete wavelet transform. We have used discrete wavelet transform in this paper. Discrete wavelet transform maps a function of a continuous variable into a sequence of coefficients. If the function being expanded is a sequence of numbers, like samples of a continuous function f (x, y), the resulting coefficients are called the discrete wavelet transform of f (x, y) [1-2]. Discrete wavelet transform of an image f (x, y) of size M N is defined as

79

International Journal of Computer Science and Engineering Research and Development (IJCSERD), ISSN 22489363 (Print), ISSN 2248-9371 (Online) Volume 1, Number 2, May-October (2011)

W ( j0 , m , n ) =
i

1 MN
1 MN

M 1 N 1

f ( x, y )
x =0 y = 0
M 1 N 1

j0 ,m ,n

( x, y )

.(1)

W ( j , m , n ) =
for j j 0 and

x =0 y =0

f ( x , y )

i j ,m ,n

( x, y )

, i = {H, V, D} . (2)

f ( x, y ) =

1 MN
+ 1 MN

W ( j
m n

, m , n ) j0 , m , n ( x , y )
i i j ,m ,n

W ( j , m , n )
i = H ,V , D j = j0 m n

( x, y )
....(3)

where, j0 is an arbitrary starting scale and

j , m , n ( x , y ) = 2 2 ( 2 j x m, 2 j y n )
and

, , i = {H, V, D}

.(4) .(5)

i j , m, n

( x, y) = 2 2 i (2 j x m,2 j y n)

where index i identifies the directional wavelets that assumes the values H,V and D. Wavelets measure functional variations such as intensity or gray-level variations for images along different directions. Directional wavelets are H , V and D . H measures variations along columns (like horizontal edges), V measures variations along rows (like vertical edges) and D measures variations along diagonals [1-2]. Eq. (4) and Eq. (5) define the scaled and translated basis functions. f (x, y), j ,m ,n ( x, y ) and ij ,m ,n ( x, y ) are functions of the discrete variables x = 0, 1, 2, 3, , M-1 and y = 0, 1, 2, 3, ..., N-1. The coefficients defined in Eq. (1) and Eq. (2) are usually called approximation and detail coefficients, respectively. W ( j0 , m, n) coefficients define an approximation of f (x, y) at scale j0 . Wi ( j, m, n) coefficients add horizontal, vertical and diagonal details for scales j j 0 . We normally let j0 = 0 and select N = M = 2 j so that j = 0, 1, 2, 3, , J - 1 and m, n = 0, 1, 2, 3, , 2 j - 1 [1-2]. Eq. (3) shows that f (x, y) is obtained via the inverse discrete wavelet transform for given W and Wi of Eq. (1) and Eq. (2). Discrete wavelet transform can be implemented using digital-filters and down-samplers [1-2]. The block diagram of multiresolution using discrete wavelet transform is shown in the Fig. 5.
80

International Journal of Computer Science and Engineering Research and Development (IJCSERD), ISSN 22489363 (Print), ISSN 2248-9371 (Online) Volume 1, Number 2, May-October (2011)

We have used MATLAB toolbox for multiresolution by discrete wavelet transform. Discrete wavelet transform is applied three times successively on each dilated character image generated in the section B, and finally the reduced character image is captured into a bounding box using MATLAB toolbox and then, resized into 21 by 15 pixels image. The effect of multiresolution technique on a character obtained in section B is given in the following figure.

Dilated Image After Slant Correction

Character Image After Multiresolution

Character Image Captured into Bounding Box

Resized Into 21 by 15

Figure 5 Character image after applying multiresolution technique and captured into bounding box. D. Classification and Recognition In this section, Euclidean distance metric is used to classify the input pattern vectors. Pattern vectors generated in the section C are grouped into 26 classes based on their similar properties. Each class contains pattern vectors, mean vector of each class is computed. We prepare a matrix whose rows are these mean vectors. Distance from input pattern vector to each mean vector is computed and minimum distance shows the class of input pattern vector to which it belongs [1418]. Suppose, pi shows the pattern vector , w j shows the pattern class , m j shows the mean vector of jth class and xi shows the component of pattern vector. Mean vector is used as the representative of the class of vectors :

mj =

1 Nj

p
pi w j

, j = 1, 2, 3, , W and i = 1, 2, 3, , N j

Where, N j is the number of pattern vectors in the class w j and W is the number of pattern classes. Euclidean distance classifier is defined as :

D j ( x) = x m j

( x m j )T ( x m j )

, j = 1, 2, 3, , W

81

International Journal of Computer Science and Engineering Research and Development (IJCSERD), ISSN 22489363 (Print), ISSN 2248-9371 (Online) Volume 1, Number 2, May-October (2011)

Where, x is the input pattern vector and D j (x ) is the distance from x to mean vector of the jth class. Input pattern vector x will belong to class wj if D j (x ) is the smallest distance [4-5].
III. EXPERIMENTAL RESULTS

20 samples (520 characters) were collected from 20 persons of different age groups, 26 from each. First, 15 samples (390 characters) were used to train the proposed HCR system and next 5 samples (130 characters) were used to test the proposed HCR system. An analysis of experimental results has been performed and shown in the table given below.

TABLE I. RESULTS SHOWING THE AVERAGE RECOGNITION ACCURACY. THE ACCURACY MEASURED WAS 90%.

Test Data Set (A Z) 35 by 25 Pixels 35 by 25 Pixels 35 by 25 Pixels 28 by 20 Pixels 28 by 20 Pixels 21 by 15 Pixels 21 by 15 Pixels 14 by 10 Pixels

Level of Multiresolution 1 2 3 2 3 2 3 3
IV. CONCLUSION

Average Recognition Accuracy (%) 86.15 86.92 85.38 87.69 86.92 86.92 90.00 89.23

From the above Table, we observe that the recognition accuracy is high for a particular level of multiresolution and appropriate resolution of input character images. When average recognition accuracy is optimal (maximum) for a particular level of multiresolution and appropriate resolution of character images, here it is 90%, then any further increment in the level of multiresolution results in the decrease of the average recognition accuracy. The resolution of input character image and the level of multiresolution are dependent upon each other so the challenge is to find appropriate relationship between them. Work is going on , we are trying to improve the recognition accuracy using some other methods and techniques.
REFERENCES

[1] Rafael C. Gonzalez and Richard E. Woods , 2002, Digital Image Processing, Pearson Education, South Asia. [2] Stephane Mallat, 2008, A Wavelet Tour of Signal Processing, Academic Press, Burlington. [3] Sergios Theodoridis and Konstantinos Koutroumbas, 2010, An Introduction to Pattern Recognition : A MATLAB Approach, Elsevier, Burlington.

82

International Journal of Computer Science and Engineering Research and Development (IJCSERD), ISSN 22489363 (Print), ISSN 2248-9371 (Online) Volume 1, Number 2, May-October (2011)

[4] Richard O. Duda, Peter E. Hart and David G. Stork, 1993, Pattern Classification, WilyIndia, UK. [5] Rafael C. Gonzalez and Richard E. Woods , 2004, Digital Image Processing using MATLAB, Pearson Education, South Asia. [6] Rejean Plamondon and Sargur N. Srihari, 2000, On-Line and Off-Line Handwriting Recognition: A Comprehensive Survey, IEEE Transactions on Pattern Analysis and Machine Intelligence, 22(1) , pp. (63-84). [7] Patrick Wunsch and Andrew F. Laine, 1995, Wavelet Descriptors for Multiresolution Recognition of Handprinted Characters, Pattern Recognition, 28(8), pp. (1237-1249). [8] Diego J. Romero, Leticia M. Seijas, Ana M. Ruedin, 2007, Directional Continuous Wavelet Transform Applied to Handwritten Numerals Recognition Using Neural Networks, JCS, 7(1), pp. (66-71). [9] G.Y. Chen, T.D. Bui, A. Krzyzak, 2003, Contour-based handwritten numeral recognition using multiwavelets and neural networks, Pattern Recognition, 36, pp. (1597-1604). [10] Seong-Whan Lee, Chang-Hun Kim, Hong Ma and Yuan Y. Tang , 1996, Multiresolution Recognition of Unconstrained Handwritten Numerals with Wavelet Transform and Multilayer Cluster Neural Network, Pattern Recognition, 29(12), pp. (1953- 1961). [11] Y. Mizukami, 1998, A handwritten Chinese character recognition system using hierarchical displacement extraction based on directional features, Pattern Recognition Letters, 19, pp. (595-604). [12] J.C. Lee, T.J. Fong and Y.F. Chang, 2009, Feature Extraction for Handwritten Chinese Character Recognition Using X-Y Graphs Decomposition and Haar Wavelet, IEEE International Conference on Signal and Image Processing Applications, pp. (10-14). [13] Amir Mowlaei, Karim Faez and Abolfazl T Haghighat, 2002, Feature Extraction with Wavelet Transform for Recognition of Isolated Handwritten Farsi or Arabic Characters and Numerals, IEEE DSP, pp. (923-926). [14] Dharamveer Sharma & Deepika Gupta, 2010, Isolated Handwritten Digit Recognition using Adaptive Unsupervised Incremental Learning Technique, International Journal of Computer Applications, 7(4), pp. (27-33). [15] J. Mantas,1986, An Overview of Character Recognition Methodologies, Pattern Recognition, 19(6), pp. (425-430). [16] U. Pal, B.B. Chaudhari, 2004, Indian Script Character Recognition: a survey, Pattern Recognition, 37, pp. (1887-1899). [17] Nafiz Arica and Fatos T. Yarman-Vural, 2001, An Overview of Character Recognition Focused on Off-Line Handwriting, IEEE Transactions on Systems, Man, and CyberneticsPart C: Applications and Reviews, 31(2), pp. (216-233). [18] Trevor Hastie and Patrice Y Simard, 1998, Metrics and Models for Handwritten Character Recognition, Statistical Science, 13(1), pp. (54-65).

83