Author Topic: Handwritten Character Recognition Using Neural Network  (Read 3675 times)

0 Members and 1 Guest are viewing this topic.


  • Newbie
  • *
  • Posts: 48
  • Karma: +0/-0
    • View Profile
Handwritten Character Recognition Using Neural Network
« on: April 23, 2011, 09:30:27 am »
Author : Chirag I Patel, Ripal Patel, Palak Patel
International Journal of Scientific & Engineering Research, IJSER - Volume 2, Issue 4, April-2011
ISSN 2229-5518
Download Full Paper -

Abstract— Objective is this paper is recognize the characters in a given scanned documents and study the effects of changing the Models of ANN. Today Neural Networks are mostly used for Pattern Recognition task. The paper describes the behaviors of different Models of Neural Network used in OCR. OCR is widespread use of Neural Network. We have considered parameters like number of Hidden Layer, size of Hidden Layer and epochs. We have used Multilayer Feed Forward network with Back propagation. In Preprocessing we have applied some basic algorithms for segmentation of characters, normalizing of characters and De-skewing. We have used different Models of Neural Network and applied the test set on each to find the accuracy of the respective Neural Network.

Index Terms— Optical Character Recognition,  Artificial Nueral Network, Backpropogation Network, Skew Detection. 

Such software’s are useful when we want to convert our Hard copies into soft copies. Such software’s reduces almost 80% of the conversion work while still some verification is always required.
Optical character recognition, usually abbreviated to OCR, involves computer software designed to translate images of typewritten text (usually captured by a scanner) into machine-editable text, or to translate pictures of characters into a standard encoding scheme representing them in (ASCII or Unicode). OCR began as a field of research in artificial intelligence and machine vision. Though academic research in the field continues, the focus on OCR has shifted to implementation of proven techniques[4].

Pattern recognition is extremely difficult to automate. Animals recognize various objects and make sense out of large amount of visual information, apparently requiring very little effort. Simulating the task performed by ani-mals to recognize to the extent allowed by physical limi-tations will be enormously profitable for the system. This necessitates study and simulation of Artificial Neural Network. In Neural Network, each node perform some simple computation and each connection conveys a signal from one node to another labeled by a number called the “connection strength” or weight indicating the extent to which signal is amplified or diminished by the connection.

Fig. 1 A simple Neuron ( Download Full Paper for Fig. 1 )

Different choices for weight results in different functions are being evaluated by the network. If in a given network whose weight are initial random and given that we know the task to be accomplished by the network , a learning algorithm must be used to determine the values of the
weight that will achieve the desired task. Learning Algo-rithm qualifies the computing system to be called Artifi-cial Neural Network. The node function was predeter-mined to apply specific function on inputs imposing a fundamental limitation on the capabilities of the network.

Typical pattern recognition systems are designed using two pass. The first pass is a feature extractor that finds features within the data which are specific to the task being solved (e.g. finding bars of pixels within an image for character recognition). The second pass is the classifier, which is more general purpose and can be trained using a neural network and sample data sets. Clearly, the feature extractor typically requires the most design effort, since it usually must be hand-crafted based on what the application is trying to achieve.

One of the main contributions of neural networks to pat-tern recognition has been to provide an alternative to this design: properly designed multi-layer networks can learn complex mappings in high-dimensional spaces without requiring complicated hand-crafted feature extractors. Thus, rather than building complex feature detection algorithms, this paper focuses on implementing a standard backpropagation neural network. It also encapsulates the Preprocessing that is required for effective.

2.1 Backpropogation
Backpropagation was created by generalizing the Wi-drow-Hoff learning rule to multiple-layer networks and nonlinear differentiable transfer functions. Input vectors and the corresponding target vectors are used to train a network until it can approximate a function, associate input vectors with specific output vectors, or classify input vectors in an appropriate way as defined by you. Networks with biases, a sigmoid layer, and a linear out-put layer are capable of approximating any function with a finite number of discontinuities.

By analyzing the OCR we have found some parameter which affects the accuracy of OCR system [1][5]. The pa-rameters listed in these papers are skewing, slanting, thickening, cursive handwriting, joint characters. If all these parameters are taken care in the preprocessing phase then overall accuracy of the Neural Network would increase.

Initially we are making the Algorithm of Character Ex-traction. We are using MATLAB as tool for implementing the algorithm. Then we design neural network, we need to have a Neural Network that would give the optimum results [2]. There is no specific way of finding the correct model of Neural Network. It could only be found by trial and error method. Take different models of Neural Network, train it and note the output accuracy.
There are basically two main phases in our Paper:

Preprocessing and Character Recognition .
In first phase we have are preprocessing the given scanned document for separating   the Characters from it and normalizing each characters. Initially we specify an input image file, which is opened for reading and preprocessing. The image would be in RGB format (usually) so we convert it into binary format. To do this, it converts the input image to grayscale format (if it is not already an intensity image), and then uses threshold to convert this grayscale image to binary i.e all the pixels above certain threshold as 1 and below it as 0.
Firstly we needed a method to extract a given character from the document. For this purpose we modified the graphics 8-way connected algorithm (which we call as EdgeDetection).

Read More: