Improvement On Triangle Features Based Grouping Features for Offline Digit Handwriting

Authors

  • N. A. Arbain Computational Intelligence and Technologies Lab (CIT LAB), Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, Durian Tunggal, 76100 Melaka MALAYSIA.
  • M. S. Azmi Computational Intelligence and Technologies Lab (CIT LAB), Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, Durian Tunggal, 76100 Melaka MALAYSIA.
  • A. K. Muda Computational Intelligence and Technologies Lab (CIT LAB), Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, Durian Tunggal, 76100 Melaka MALAYSIA.
  • A. R. Radzid Computational Intelligence and Technologies Lab (CIT LAB), Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, Durian Tunggal, 76100 Melaka MALAYSIA.
  • A. Tahir Computational Intelligence and Technologies Lab (CIT LAB), Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, Durian Tunggal, 76100 Melaka MALAYSIA. Department of Information & Communication Technology, Politeknik Ungku Omar, Jalan Raja Musa Mahadi, 31400 Ipoh, Perak, MALAYSIA.

Keywords:

Digit Handwriting, Support Vector Machine, Triangle Geometry, Triangle Features,

Abstract

An offline digit handwriting recognition is one of an active studied that has been explored in the field of pattern recognition. In this paper, an improvement on triangle features based grouping features is proposed. It uses to overcome the problem of processing data where the performance is slow based on time training. This problem occurred due to the huge size of the number of triangle features are used. The grouping features are focused on triangle properties of ratio and gradient where the outcome of this grouping features will produce five triangle features which are gRatio-ABC, gGradient-ABC, angle point A, angle point B and angle point C. Then, the converting process using the absolute value function is applied to increase the classification accuracies for digit dataset of IFCHDB, HODA, MNIST and BANGLA. A classifier of Support Vector Machine was used to measure the accuracies.

Downloads

Download data is not yet available.

Downloads

Published

2018-07-04

How to Cite

Arbain, N. A., Azmi, M. S., Muda, A. K., Radzid, A. R., & Tahir, A. (2018). Improvement On Triangle Features Based Grouping Features for Offline Digit Handwriting. Journal of Telecommunication, Electronic and Computer Engineering (JTEC), 10(2-7), 59–62. Retrieved from https://jtec.utem.edu.my/jtec/article/view/4420

Most read articles by the same author(s)