Digital Image Processing (DIP)

Reconstruction of Character Using Image Processing Technique

The system based document processing is one of the major trends of office automation. The goal of Optical Character Recognition (OCR) is to classify optical patterns (often contained in a digital image) corresponding to alphanumeric or other characters. Recognition of handwritten numerals is important because of its applicability in various fields like postal code recognition, phone no., check processing etc.

The recognition can be achieved by many methods such as dynamic programming, hidden Markov modeling, neural network, nearest neighbor classifier, expert system and a combination of all these techniques.

Objectives:

  • To study various existing techniques and methodologies for offline handwritten and printed digit recognition
  • To study the feature extraction and classification stages in solving any pattern recognition problem
  • To study the structural, statistical techniques for feature extraction and classification phase of Automatic Recognition of Offline Handwritten and printed English digit of various size and types
  • To study the hybrid techniques for feature extraction and classification phase of Automatic Recognition of Offline Handwritten and printed English digit of various size and types

 

Classification

  • Optical Character Recognition can be classified into
  • Online OCR
  • Offline OCR
  • Printed OCR
  • Handwritten OCR
  • Hand painted OCR

The English characters and numerals are written from left to right but only English Characters are written in cursive style shown in Fig.1.

0     1    2  3  4  5  6  7  8  9

A     B    C  D  E  F  G  H  I      J  K  

N    O    P  Q  R  S  T  U  V  W

X    Y    Z

a    b    c  d  e  f  g  h  i  j   k

m   n    o  p  q  r  s  t  u     v  w

x     y    z

 

0     1    2  3  4  5  6  7  8      9

A     B    C  D  E  F  G  H  J  K

L    M      N    O  P  Q  R  S  T  V

W   X  Y  Z

 

a    b  c  d  e  f  g  h  I  j

k  l  m  n  o  p  q  r  s  t

u  v  w  x  y  z

Fig.1 (a) Printed English Digits and Characters.

Methodology
Character Recognition is a part of a broad domain called Pattern Recognition. The Pattern Recognition problems can be solved by following steps.

  • Preprocessing
  • Feature Extraction
  • Classification
  • Recognition/Post Processing

HYPOTHESIS

It should be possible to implement character recognition and reconstruction using image processing techniques employing classification techniques like neural networks. This approach is likely to provide solutions for reconstructing correct characters to appreciable extent.

SCOPE & LIMITATIONS

English characters and digits with printed as well as  handwritten  with different style and font size to reconstruct and correct them with the help of image processing is achieved. In feature English character as well as digits printed and handwritten image in to text file

Due to handwriting variation of the different samples of same writers as well as different writers the accuracy is up to 90-95% is achieved

Leave a Reply

Your email address will not be published. Required fields are marked *