Assamese Character Recognition Using Zoning Feature
The aim of this paper is to develop a model to recognize basic Assamese characters using feed-forward neural network. The basic characters included in Assamese language are a set of numeral, a set of vowel, and a set of consonant. An algorithm has been de
- PDF / 305,396 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 15 Downloads / 199 Views
Abstract The aim of this paper is to develop a model to recognize basic Assamese characters using feed-forward neural network. The basic characters included in Assamese language are a set of numeral, a set of vowel, and a set of consonant. An algorithm has been designed to segment the line and individual character of the image and zoning features are extracted from the individual character. The network is trained by gradient descent with momentum and adaptive learning rate backpropagation training function. The network consists of two hidden layers with Sum Square Error (SSE). Finally, the unicode value of the recognized character is written in a text file. Keywords Assamese character recognition Segmentation Zoning Back propagation
Feed-forward neural network
1 Introduction In pattern recognition it is still a challenging task to produce an editable document from handwritten or machine-written characters. There are two ways to recognize characters: recognize while writing or recognize after writing. The former is called online and the latter is offline recognition technique. This experiment tries to implement offline technique. Character recognition system is used in bank for cheque processing and sorting, where the digits of cheque numbers are special font printed with magnetic ink that is prone to noise. Moreover, mobile applications are available that allows user to extract information automatically from visiting card of a person to their mobile by taking a snapshot of the card. In post office it can be
K. Medhi (&) S.Kr. Kalita Department of Computer Science, Gauhati University, Guwahati, India e-mail: [email protected] S.Kr. Kalita e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2018 A. Kalam et al. (eds.), Advances in Electronics, Communication and Computing, Lecture Notes in Electrical Engineering 443, https://doi.org/10.1007/978-981-10-4765-7_39
371
372
K. Medhi and S.Kr. Kalita
used for postal number reading and sorting. Publication house can rapidly produce editable document from old books for publishing using CR system. In India, OCR system has been developed for Bangla, Devanagari, Gurumukhi, Kannada, Malayalam, Tamil, and Telugu [1]. Thus, OCR systems for Indian scripts have just started appearing. Literature review reveals that research on Assamese language is very tiny and that is why this is an attempt to design a CR system for this language.
2 Properties of Assamese Scripts The script used in Assamese language is originated from an ancient Indian script called Brahmi [2]. Most of the characters in Assamese script have a horizontal line from where it hang, called the head line or matra, e.g., ক, ঘ, অ, আ, ই, etc. The vowel can present independently in a word or they may act as modifier with consonant. The consonant can be independent, but for pronunciation they took help of vowel. Combination of two or three consonants can produce another compound character.
3 Literature Review Sharma and Gupta [3] proposed a system with a k-nn classifier with k = 1 for
Data Loading...
