as

Construction of Co-occurrence Matrix using Gabor Wavelets for Classification of Arecanuts by Decision Trees

Suresh, M and Danti, Ajit (2012) Construction of Co-occurrence Matrix using Gabor Wavelets for Classification of Arecanuts by Decision Trees. International Journal of Applied Information Systems (IJAIS), 4 (6). pp. 33-39. ISSN 2249-0868

[img] Text
ku-art-cs-sm-ijais-122012.pdf - Published Version
Restricted to Registered users only

Download (759kB) | Request a copy
Official URL: http://www.ijais.org

Abstract

In this paper, a novel method developed for classification of arecanuts based on texture features. A Gabor response co-occurrence matrix (GRCM) is constructed analogous to Gray Level co-occurrence matrix (GLCM). Classification is done using kNN and Decision Tree (DT) classifier based on GRCM features. There are twelve Gabor filters are designed (Four orientations and three scaling) to capture variations in lighting conditions. Results are compared with kNN classifier and found better results. Splitting rules for growing decision tree that are included are gini diversity index(gdi), twoing rule, and entropy. Proposed approach is experimented on large data set using cross validation and found good success rate in decision tree classifier.

Item Type: Article
Uncontrolled Keywords: Image Processing, Pattern Recognition, Classification.
Subjects: 000 Computer Science, Information and General work > 004 Computer Science
Divisions: Department of Computer Science
Depositing User: Unnamed user with email bs@kuvempu.ac.in
Date Deposited: 26 Aug 2017 12:39
Last Modified: 27 Aug 2017 05:14
URI: http://kuls-ir.kuvempu.ac.in/id/eprint/1721

Actions (login required)

View Item View Item