Open AccessComparison of Feature Extraction Techniques for Pattern Classification
Author(s) -
Binu P. Chacko
Publication year - 2021
Publication title -
international journal for research in applied science and engineering technology
Language(s) - English
Resource type - Journals
ISSN - 2321-9653
DOI - 10.22214/ijraset.2021.36214
Subject(s) - pattern recognition (psychology) , malayalam , discriminative model , artificial intelligence , computer science , feature extraction , classifier (uml) , division (mathematics) , support vector machine , point (geometry) , feature (linguistics) , mathematics , linguistics , philosophy , geometry , arithmetic
Pattern recognition is a challenging task in research field for the last few decades. Many researchers have worked on areas such as computer vision, speech recognition, document classification, and computational biology to tackle complex research problems. In this article, a pattern recognition problem for handwritten Malayalam character is presented. This system goes through two different stages of HCR namely, feature extraction and classification. Three feature extraction techniques – wavelet transform, zoning, division point – are used in this study. Among these, division is point is able to show best discriminative power using SVM classifier. All the experiments are conducted on size normalized and binarized images of isolated Malayalam characters.