A Hybrid Neural Approach For

Published on February 2017 | Categories: Documents | Downloads: 27 | Comments: 0 | Views: 312
of 3
Download PDF   Embed   Report

Comments

Content

INTRODUCTION
Object character recognition (OCR) systems can contribute tremendously to the advancement of the automation process and can improve the interaction between man and machine in many applications, including office automation, check verification and a large variety of banking, business and data entry applications. The recognition rate of OCR (Optical Character Recognition)systems often decreases fordegraded input digit images,such as stroke-connected, stroke-broken and noisy images. This can be caused by a lot of reasons, such asthe poor quality of the source paper material, the bad quality of printing, the wrong threshold selected for scanning. The recognition process of an OCR system can be divided into several steps like, preprocessing, segmentation, feature extraction and classification. History of OCR, has been discussed by Shunji Mori et.al. Toshiyuki Sakai et.al described documentin formation processing both on hardware and software.FaisalFarooq et. al. defined the role of preprocessing in the handwritten recognition. They say that to improve the readability and the automatic recognition of handwritten document images, preprocessing steps are imperative. Hongwei Konget.al.presented unified method for preprocessing binary text image based on mathematical morphology while S.N. Nawaz et. al Attaullan Khawaja et.al define the basic use of preprocessing in the character recognition. Yungang Zhang and Changshui Zhang efine enhancement technique for image preprocessing. They used normalization and enhancement technique for preprocessing. M. Sarfraz et.al proposed a new technique for skew estimation of image document. Pei- Yung Hsiao. et.alin the paper “Generic 2-D Gaussian Smoothing Filter for Noisy Image Processing” define the use of Gaussian filter in OCR. G. Deng and L.W. Cahill propose an adaptive Gaussian filtering algorithm in which the filter variance is adapted to both the noise characteristics and the local variance of the signal. H. Wehbi et. al. provide two segmentation modules, the first one is to isolate the word drawings within a sentence, and the other one is to separate numeral characters and capital letters from a mixed writing prior to their recognition. Xiaoping Li et. al. lso proposedthe connected domain analysis method, segmentation projection histogram method and shelling projection method for segmentation of digits. M. Blumenstein et. al. Introduced new feature extraction technique investigated whose research was based on the calculation and location of transition features from background to foreground pixels in the vertical and horizontal directions. Brijesh 1

Verma et.al proposed feature extraction technique that can be classified into eight modules such as dehooking, extract feature points, stroke extracting, calculate PEN UP, extract zones and directions of start point and end point, extract changes in writing direction, calculate height width ratio and extract zone information which creates a global feature vector and uses a back-propagation neural network based classifier. S.V. Rajashekararadhya and Vanaja Ranjan P propose a zone-based hybrid feature extraction system. R.J. Ramteke et.al present an experimental evaluation of the effectiveness of various techniques based upon moment invariants. Douglas Kozlay describes a feature extraction technique for the next generation of optical character readers. Hermineh Sanossian tells us that extract relevant features from numeral images reduce the complexity of recognition. He introduces the new method in his paper. He finds different features, features due to boundary distances in a segment the pixel densities in a segment, and line distances from centroid in a segment. Chongliang Zhong et.al introduce the 13-point feature of skeleton method. There are two steps that they take. One is to find the skeleton of the character and the other is to extract the 13-point feature of skeleton. S.V. Rajashekararadhya et.al present Zone andDistance metric based feature extraction system. Abdur Rahim Md. Forkan et.al. describe in the paper that how multi layered feed forward artificial neural network for classification of character and position of white pixels of the images considered as noise and high pass filtering is used to remove this noise from input image. Zaheer Ahmad et.al developed a system consists of two main modules segmentation and classification. Ping Zhang, Lihui Chen, Alex C Kot proposed a hybrid neural network and tree classification system for handwritten numeral recognition. Christopher L. Scofield, Lannie Kenton, Jung-Chou Chang proposed a multiple neural network system (MNNS) for image-based character recognition. Based on artificial neural network, digital image processing, and features extraction theory, Huang Hanmin Huang et.al analyzed BP network’s affect and presented its improving solutions. Hermineh Y.Y. Sanossian described an optical character recognition system which uses multilayer perceptron classifier. Michael D. Garris defines fourier descriptors, moment invariants, and other boundary features. Adrian Lim Hooi Jin et.al say that all the processed fingerprint image are used as an input to the back propagation neural network designed to perform the training process. Thus, a large number of researchers have published paper on this area each has own advantage and

2

disadvantage. In fact the need of hour is to develop a digit recognition system for improving recognition accuracy. So in the present work back propagation neural network based digit recognition system has been proposed. ABOUT HARDWARE & SOFTWARE REQUIREMENTS Software Requirements . • • • Operating system Front End Coding Language : Windows XP Professional : Microsoft Visual Studio .Net 2008 : Visual C# .Net

Hardware Requirements • • • • SYSTEM HARD DISK MOUSE RAM : Pentium III 700 MHz : 40 GB. : Logitech. : 512 MB.

3

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close