Classification of images for automatic textual annotation: A review of techniques

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Annotating images with text is one of the approaches to represent semantic meaning of images. Automatic classification of images into various semantic categories is one of the many steps required to perform the automatic textual annotation as exhibited in many research in this area. However, little or none researchers in this area do provide detail evaluation for selecting the best or suitable technique for performing image classification. The majority of the researchers mainly select any of the available machine learning technique and apply it as part of their proposed approaches and algorithms. In this study, six techniques were reviewed, which are SVM, Multilayer Perceptron, Bagging, DECORATE, C4.5 Decision Tree and Random Forest using 429 Flickr images relating to Malaysian tourism. Image feature extraction using 3D Colour Histogram with 64 and 216 bins were done. The results show that DECORATE has the best accuracy.

Original languageEnglish
Pages (from-to)760-767
Number of pages8
JournalJournal of Applied Sciences
Volume13
Issue number6
DOIs
Publication statusPublished - 2013

Fingerprint

Semantics
Image classification
Bins
Multilayer neural networks
Decision trees
Learning systems
Feature extraction
Color

Keywords

  • Automatic textual annotation
  • DECORATE
  • Image annotation
  • Image classification
  • Machine Learning
  • Semantic

ASJC Scopus subject areas

  • General

Cite this

Classification of images for automatic textual annotation : A review of techniques. / Ghazali, Juzlinda; Mohd Noah, Shahrul Azman; Zakaria, Lailatul Qadri.

In: Journal of Applied Sciences, Vol. 13, No. 6, 2013, p. 760-767.

Research output: Contribution to journalArticle

@article{8131754ebee5453c9a7b03313ef725b3,
title = "Classification of images for automatic textual annotation: A review of techniques",
abstract = "Annotating images with text is one of the approaches to represent semantic meaning of images. Automatic classification of images into various semantic categories is one of the many steps required to perform the automatic textual annotation as exhibited in many research in this area. However, little or none researchers in this area do provide detail evaluation for selecting the best or suitable technique for performing image classification. The majority of the researchers mainly select any of the available machine learning technique and apply it as part of their proposed approaches and algorithms. In this study, six techniques were reviewed, which are SVM, Multilayer Perceptron, Bagging, DECORATE, C4.5 Decision Tree and Random Forest using 429 Flickr images relating to Malaysian tourism. Image feature extraction using 3D Colour Histogram with 64 and 216 bins were done. The results show that DECORATE has the best accuracy.",
keywords = "Automatic textual annotation, DECORATE, Image annotation, Image classification, Machine Learning, Semantic",
author = "Juzlinda Ghazali and {Mohd Noah}, {Shahrul Azman} and Zakaria, {Lailatul Qadri}",
year = "2013",
doi = "10.3923/jas.2013.760.767",
language = "English",
volume = "13",
pages = "760--767",
journal = "Journal of Applied Sciences",
issn = "1812-5654",
publisher = "Asian Network for Scientific Information",
number = "6",

}

TY - JOUR

T1 - Classification of images for automatic textual annotation

T2 - A review of techniques

AU - Ghazali, Juzlinda

AU - Mohd Noah, Shahrul Azman

AU - Zakaria, Lailatul Qadri

PY - 2013

Y1 - 2013

N2 - Annotating images with text is one of the approaches to represent semantic meaning of images. Automatic classification of images into various semantic categories is one of the many steps required to perform the automatic textual annotation as exhibited in many research in this area. However, little or none researchers in this area do provide detail evaluation for selecting the best or suitable technique for performing image classification. The majority of the researchers mainly select any of the available machine learning technique and apply it as part of their proposed approaches and algorithms. In this study, six techniques were reviewed, which are SVM, Multilayer Perceptron, Bagging, DECORATE, C4.5 Decision Tree and Random Forest using 429 Flickr images relating to Malaysian tourism. Image feature extraction using 3D Colour Histogram with 64 and 216 bins were done. The results show that DECORATE has the best accuracy.

AB - Annotating images with text is one of the approaches to represent semantic meaning of images. Automatic classification of images into various semantic categories is one of the many steps required to perform the automatic textual annotation as exhibited in many research in this area. However, little or none researchers in this area do provide detail evaluation for selecting the best or suitable technique for performing image classification. The majority of the researchers mainly select any of the available machine learning technique and apply it as part of their proposed approaches and algorithms. In this study, six techniques were reviewed, which are SVM, Multilayer Perceptron, Bagging, DECORATE, C4.5 Decision Tree and Random Forest using 429 Flickr images relating to Malaysian tourism. Image feature extraction using 3D Colour Histogram with 64 and 216 bins were done. The results show that DECORATE has the best accuracy.

KW - Automatic textual annotation

KW - DECORATE

KW - Image annotation

KW - Image classification

KW - Machine Learning

KW - Semantic

UR - http://www.scopus.com/inward/record.url?scp=84880157591&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84880157591&partnerID=8YFLogxK

U2 - 10.3923/jas.2013.760.767

DO - 10.3923/jas.2013.760.767

M3 - Article

AN - SCOPUS:84880157591

VL - 13

SP - 760

EP - 767

JO - Journal of Applied Sciences

JF - Journal of Applied Sciences

SN - 1812-5654

IS - 6

ER -