عنوان البحث(Papers / Research Title)
Numeral Recognition Using Statistical Methods
الناشر \ المحرر \ الكاتب (Author / Editor / Publisher)
ندى عبد الله رشيد الجبوري
Citation Information
ندى,عبد,الله,رشيد,الجبوري , Numeral Recognition Using Statistical Methods , Time 7/15/2011 4:29:38 PM : كلية التربية الاساسية
وصف الابستركت (Abstract)
We Proposed Algorithm
الوصف الكامل (Full Abstract)
Numeral Recognition Using Statistical Methods Comparison Study
Huda A. Rasheed Nada A. Rasheed
AL- Mustansiriyah University Babylon University
College of sciences College of Basic Education
2011
Abstract
The area of character recognition has been received a considerable attention by researchers all over the world during the last three decades. However, in this research explores best sets of feature extraction techniques and studies the accuracy of well-known classifiers for Arabic numeral using the Statistical styles in two methods and comparison study between them. First methods Linear Discriminant function, that are yield results with accuracy as high as 90% of original grouped cases correctly classified. Second method we proposed algorithm, the results show the efficiency of the proposed algorithms, where it is found to achieve recognition accuracy of 92.9% and 91.4%. This is providing efficient is more than the first method.
Publication: Baghdad / Science JournalVolume : 8(1)2011Number: 19Starting page: 183Ending page: 188Keywords: Numeral, Recognition, Discriminant Nada A. Rasheed
Introduction
Document image analysis 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 [1].
Character recognition is a long-standing, fundamental problem in pattern recognition. It has been the subject of a considerable number of studies and serves many useful applications. Of the two major issues in character recognition, character shape representation and category assignment, classification has been, by and large, the subject of most studies. These studies assumed that descriptions of shape by basic characteristics such as curvature, tangents, and transform coefficients, are sufficiently expressive to justify focusing on classification. However, a good representation is as important as a good classifier for high performance [2].
There are two types of character recognition systems: on-line and off-line systems. Each system has its own algorithms and methods. The main difference between them is that in an on-line system the recognition is performed in the time of writing while the off-line recognition is performed after the writing is completed [3].
The recognition of hand4written numeral characters has been a topic widely studied during the recent decades because of both its theoretical value in pattern recognition and its numerous possible applications [4], one such area is the reading of courtesy amounts on bank checks. This application has been very popular in handwriting recognition research, due to the availability of relatively inexpensive CPU power, and the possibility to reduce considerably the manual effort involved in this task. Another application is the reading of postal zip codes in addresses written or typed on envelopes. The former is more difficult than the latter due to a number of differences in the nature of the handwritten material. For example, bank checks systems [5].
Pattern recognition systems typically involve two steps: feature extraction in which appropriate representation of pattern are developed and classification in which decision rules for separating pattern classes are defined. There are indeed as many possible features as the ways characters are written. These features can be classified into two major categories: statistical and structural features. In the statistical approaches the input pattern is characterized by a set of N features and its description is achieved my means of a feature vector belonging to an N-dimensional space. On the other hand, in structural approaches it is assumed that the pattern to be recognized can be decomposed into simpler components (called primitive) and then described in terms of simple appropriate attributes of primitives and their topological relations [6].
Note: Click on the PDF icon below for downloading the paper.
تحميل الملف المرفق Download Attached File
|
|