Feature Extraction Method for Classification of Approved Halal Logo in Malaysia using Fractionalized Principle Magnitude  

Saipullah, Khairul Muzzammil (2013) Feature Extraction Method for Classification of Approved Halal Logo in Malaysia using Fractionalized Principle Magnitude. Engineering Management Reviews (EMR), 2 (2). 36 -44 . ISSN 2326-5884

[img] PDF
EMR017.pdf - Published Version
Restricted to Registered users only

Download (1MB) | Request a copy

Abstract

This  paper  presents  new  feature  extractors  called  the Fractionalized  Principle  Magnitude  (FPM)  that  is  evaluated in  the  classification  of  approved  Halal  logo  with  respect  to classification  accuracy  and  time  consumptions.  Feature  can be classified into two groups  that are global and local feature.  In  this  study,  several  feature  extractors  have been  compared with  the  proposed  method  such  as  histogram  of  gradient (HOG),  Hu  moment,  Zernike  moment  and  wavelet  co‐occurrence  histogram  (WCH).  The  experiments  are conducted  on  50  different  approved  Halal  logos.  The  result shows  that  proposed  FPM  method  achieves  the  highest accuracy  with  90.4%  whereas  HOG,  Zernike  moment,  WCH and  Hu  moment  achieve  75.2%,  64.4%,  47.2%  44.4%  of accuracies,  respectively.  Furthermore,  two  other  databases also  have  been  used  that  is  traffic  sign  and  Outex  database. The  accuracy  performance  and  classification  time  are compared with FPM and other method. 

Item Type: Article
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
Divisions: Faculty of Electronics and Computer Engineering > Department of Computer Engineering
Depositing User: Engr. Khairul Muzzammil Saipullah
Date Deposited: 12 Jul 2013 09:14
Last Modified: 28 May 2015 03:57
URI: http://eprints.utem.edu.my/id/eprint/8557
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item