Kurdish Handwritten Classification Based on Human Personality Analysis

Authors

  • Ahmed Abdullah Ahmed Faculty of Engineering & Science Qaiwan International University (QIU) Raparin, Sulaymaniyah, Kurdistan Region-Iraq

DOI:

https://doi.org/10.25007/ajnu.v10n1a862

Keywords:

Grouping, writer identification, feature extraction, feature combination, database

Abstract

This study proposes a holistic technique of classifying Kurdish handwritten text documents, involving several stages. The first stage entails the sectioning of Kurdish handwritten document images into lines, words, and characters and the second stage entails the obtaining of structural and statistical features from the sectioned parts that are often utilized in human personality analysis for the examination of feature behavior and effectiveness. This is done by combining the entire potential outcomes to determine the significant features set. The third stage entails the use of F-Measure to evaluate the extracted features performance and their combination in various relationship methods, individually and in groups. The last stage entails the actual experiment using the standard KRDOH dataset of the Kurdish handwritten text, containing 1076 volunteers’ samples of different ages, genders, and education levels in a cumulative 4304 manuscripts consisting of 4 contributed pages by each writer. Based on the results obtained from multiple runs of an individual clustering method of each distance measure, a good set of features generally deliver significantly enhanced clustering of handwritten structures.

Downloads

Download data is not yet available.

Published

2021-03-08

How to Cite

Ahmed, A. A. (2021). Kurdish Handwritten Classification Based on Human Personality Analysis. Academic Journal of Nawroz University, 10(1), 174–182. https://doi.org/10.25007/ajnu.v10n1a862

Issue

Section

Articles