Utilizing context information to enhance content-based image classification

Qiusha Zhu, Lin Lin, Mei-Ling Shyu, Dianting Liu

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Traditional image classification relies on text information such as tags, which requires a lot of human effort to annotate them. Therefore, recent work focuses more on training the classifiers directly on visual features extracted from image content. The performance of content-based classification is improving steadily, but it is still far below users' expectation. Moreover, in a web environment, HTML surrounding texts associated with images naturally serve as context information and are complementary to content information. This paper proposes a novel two-stage image classification framework that aims to improve the performance of content-based image classification by utilizing context information of web-based images. A new TF*IDF weighting scheme is proposed to extract discriminant textual features from HTML surrounding texts. Both content-based and context-based classifiers are built by applying multiple correspondence analysis (MCA). Experiments on web-based images from Microsoft Research Asia (MSRA-MM) dataset show that the proposed framework achieves promising results.

Original languageEnglish (US)
Title of host publicationMultimedia Data Engineering Applications and Processing
PublisherIGI Global
Pages114-130
Number of pages17
ISBN (Print)9781466629417, 1466629401, 9781466629400
DOIs
StatePublished - Feb 28 2013

Fingerprint

Image classification
HTML
Classifiers
World Wide Web
Experiments

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Zhu, Q., Lin, L., Shyu, M-L., & Liu, D. (2013). Utilizing context information to enhance content-based image classification. In Multimedia Data Engineering Applications and Processing (pp. 114-130). IGI Global. https://doi.org/10.4018/978-1-4666-2940-0.ch006

Utilizing context information to enhance content-based image classification. / Zhu, Qiusha; Lin, Lin; Shyu, Mei-Ling; Liu, Dianting.

Multimedia Data Engineering Applications and Processing. IGI Global, 2013. p. 114-130.

Research output: Chapter in Book/Report/Conference proceedingChapter

Zhu, Q, Lin, L, Shyu, M-L & Liu, D 2013, Utilizing context information to enhance content-based image classification. in Multimedia Data Engineering Applications and Processing. IGI Global, pp. 114-130. https://doi.org/10.4018/978-1-4666-2940-0.ch006
Zhu Q, Lin L, Shyu M-L, Liu D. Utilizing context information to enhance content-based image classification. In Multimedia Data Engineering Applications and Processing. IGI Global. 2013. p. 114-130 https://doi.org/10.4018/978-1-4666-2940-0.ch006
Zhu, Qiusha ; Lin, Lin ; Shyu, Mei-Ling ; Liu, Dianting. / Utilizing context information to enhance content-based image classification. Multimedia Data Engineering Applications and Processing. IGI Global, 2013. pp. 114-130
@inbook{366ca024ad7d46fdb22236006208a1ae,
title = "Utilizing context information to enhance content-based image classification",
abstract = "Traditional image classification relies on text information such as tags, which requires a lot of human effort to annotate them. Therefore, recent work focuses more on training the classifiers directly on visual features extracted from image content. The performance of content-based classification is improving steadily, but it is still far below users' expectation. Moreover, in a web environment, HTML surrounding texts associated with images naturally serve as context information and are complementary to content information. This paper proposes a novel two-stage image classification framework that aims to improve the performance of content-based image classification by utilizing context information of web-based images. A new TF*IDF weighting scheme is proposed to extract discriminant textual features from HTML surrounding texts. Both content-based and context-based classifiers are built by applying multiple correspondence analysis (MCA). Experiments on web-based images from Microsoft Research Asia (MSRA-MM) dataset show that the proposed framework achieves promising results.",
author = "Qiusha Zhu and Lin Lin and Mei-Ling Shyu and Dianting Liu",
year = "2013",
month = "2",
day = "28",
doi = "10.4018/978-1-4666-2940-0.ch006",
language = "English (US)",
isbn = "9781466629417",
pages = "114--130",
booktitle = "Multimedia Data Engineering Applications and Processing",
publisher = "IGI Global",

}

TY - CHAP

T1 - Utilizing context information to enhance content-based image classification

AU - Zhu, Qiusha

AU - Lin, Lin

AU - Shyu, Mei-Ling

AU - Liu, Dianting

PY - 2013/2/28

Y1 - 2013/2/28

N2 - Traditional image classification relies on text information such as tags, which requires a lot of human effort to annotate them. Therefore, recent work focuses more on training the classifiers directly on visual features extracted from image content. The performance of content-based classification is improving steadily, but it is still far below users' expectation. Moreover, in a web environment, HTML surrounding texts associated with images naturally serve as context information and are complementary to content information. This paper proposes a novel two-stage image classification framework that aims to improve the performance of content-based image classification by utilizing context information of web-based images. A new TF*IDF weighting scheme is proposed to extract discriminant textual features from HTML surrounding texts. Both content-based and context-based classifiers are built by applying multiple correspondence analysis (MCA). Experiments on web-based images from Microsoft Research Asia (MSRA-MM) dataset show that the proposed framework achieves promising results.

AB - Traditional image classification relies on text information such as tags, which requires a lot of human effort to annotate them. Therefore, recent work focuses more on training the classifiers directly on visual features extracted from image content. The performance of content-based classification is improving steadily, but it is still far below users' expectation. Moreover, in a web environment, HTML surrounding texts associated with images naturally serve as context information and are complementary to content information. This paper proposes a novel two-stage image classification framework that aims to improve the performance of content-based image classification by utilizing context information of web-based images. A new TF*IDF weighting scheme is proposed to extract discriminant textual features from HTML surrounding texts. Both content-based and context-based classifiers are built by applying multiple correspondence analysis (MCA). Experiments on web-based images from Microsoft Research Asia (MSRA-MM) dataset show that the proposed framework achieves promising results.

UR - http://www.scopus.com/inward/record.url?scp=84944682488&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84944682488&partnerID=8YFLogxK

U2 - 10.4018/978-1-4666-2940-0.ch006

DO - 10.4018/978-1-4666-2940-0.ch006

M3 - Chapter

SN - 9781466629417

SN - 1466629401

SN - 9781466629400

SP - 114

EP - 130

BT - Multimedia Data Engineering Applications and Processing

PB - IGI Global

ER -