Link Anchors in Images: Is there Truth?

Robin Aly, Kevin McGuinness, Martijn Kleppe, Roeland J.F. Ordelman, Noel O'Connor, Franciska M.G. de Jong

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademic

7 Downloads (Pure)

Abstract

While automatic linking in text collections is well understood, little is known about links in images. In this work, we investigate two aspects of anchors, the origin of a link, in images: 1) the requirements of users for such anchors, e.g. the things users would like more information on, and 2) possible evaluation methods assessing anchor selection algorithms. To investigate these aspects, we perform a study with 102 users. We find that 59% of the required anchors are image segments, as opposed to the whole image, and most users require information on displayed persons. The agreement of users on the required anchors is too low (often below 30%) for a ground truth-based evaluation, which is the standard IR evaluation method. As an alternative, we propose a novel evaluation method based on improved search performance and user experience.
Original languageUndefined
Title of host publicationProceedings of the 12th Dutch Belgian Information Retrieval Workshop (DIR 2012)
Place of PublicationGhent
PublisherGhent University
Pages1-4
Number of pages4
ISBN (Print)not assigned
Publication statusPublished - 24 Feb 2012
Event12th Dutch-Belgian Information Retrieval Workshop, DIR 2012 - Ghent, Belgium
Duration: 24 Feb 201224 Feb 2012
Conference number: 12

Publication series

Name
PublisherUniversity Ghent

Workshop

Workshop12th Dutch-Belgian Information Retrieval Workshop, DIR 2012
Abbreviated titleDIR
CountryBelgium
CityGhent
Period24/02/1224/02/12

Keywords

  • METIS-286282
  • EWI-21603
  • IR-80184

Cite this

Aly, R., McGuinness, K., Kleppe, M., Ordelman, R. J. F., O'Connor, N., & de Jong, F. M. G. (2012). Link Anchors in Images: Is there Truth? In Proceedings of the 12th Dutch Belgian Information Retrieval Workshop (DIR 2012) (pp. 1-4). Ghent: Ghent University.
Aly, Robin ; McGuinness, Kevin ; Kleppe, Martijn ; Ordelman, Roeland J.F. ; O'Connor, Noel ; de Jong, Franciska M.G. / Link Anchors in Images: Is there Truth?. Proceedings of the 12th Dutch Belgian Information Retrieval Workshop (DIR 2012). Ghent : Ghent University, 2012. pp. 1-4
@inproceedings{553903cbff15425b8d9c970c2d6b0440,
title = "Link Anchors in Images: Is there Truth?",
abstract = "While automatic linking in text collections is well understood, little is known about links in images. In this work, we investigate two aspects of anchors, the origin of a link, in images: 1) the requirements of users for such anchors, e.g. the things users would like more information on, and 2) possible evaluation methods assessing anchor selection algorithms. To investigate these aspects, we perform a study with 102 users. We find that 59{\%} of the required anchors are image segments, as opposed to the whole image, and most users require information on displayed persons. The agreement of users on the required anchors is too low (often below 30{\%}) for a ground truth-based evaluation, which is the standard IR evaluation method. As an alternative, we propose a novel evaluation method based on improved search performance and user experience.",
keywords = "METIS-286282, EWI-21603, IR-80184",
author = "Robin Aly and Kevin McGuinness and Martijn Kleppe and Ordelman, {Roeland J.F.} and Noel O'Connor and {de Jong}, {Franciska M.G.}",
year = "2012",
month = "2",
day = "24",
language = "Undefined",
isbn = "not assigned",
publisher = "Ghent University",
pages = "1--4",
booktitle = "Proceedings of the 12th Dutch Belgian Information Retrieval Workshop (DIR 2012)",

}

Aly, R, McGuinness, K, Kleppe, M, Ordelman, RJF, O'Connor, N & de Jong, FMG 2012, Link Anchors in Images: Is there Truth? in Proceedings of the 12th Dutch Belgian Information Retrieval Workshop (DIR 2012). Ghent University, Ghent, pp. 1-4, 12th Dutch-Belgian Information Retrieval Workshop, DIR 2012, Ghent, Belgium, 24/02/12.

Link Anchors in Images: Is there Truth? / Aly, Robin; McGuinness, Kevin; Kleppe, Martijn; Ordelman, Roeland J.F.; O'Connor, Noel; de Jong, Franciska M.G.

Proceedings of the 12th Dutch Belgian Information Retrieval Workshop (DIR 2012). Ghent : Ghent University, 2012. p. 1-4.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademic

TY - GEN

T1 - Link Anchors in Images: Is there Truth?

AU - Aly, Robin

AU - McGuinness, Kevin

AU - Kleppe, Martijn

AU - Ordelman, Roeland J.F.

AU - O'Connor, Noel

AU - de Jong, Franciska M.G.

PY - 2012/2/24

Y1 - 2012/2/24

N2 - While automatic linking in text collections is well understood, little is known about links in images. In this work, we investigate two aspects of anchors, the origin of a link, in images: 1) the requirements of users for such anchors, e.g. the things users would like more information on, and 2) possible evaluation methods assessing anchor selection algorithms. To investigate these aspects, we perform a study with 102 users. We find that 59% of the required anchors are image segments, as opposed to the whole image, and most users require information on displayed persons. The agreement of users on the required anchors is too low (often below 30%) for a ground truth-based evaluation, which is the standard IR evaluation method. As an alternative, we propose a novel evaluation method based on improved search performance and user experience.

AB - While automatic linking in text collections is well understood, little is known about links in images. In this work, we investigate two aspects of anchors, the origin of a link, in images: 1) the requirements of users for such anchors, e.g. the things users would like more information on, and 2) possible evaluation methods assessing anchor selection algorithms. To investigate these aspects, we perform a study with 102 users. We find that 59% of the required anchors are image segments, as opposed to the whole image, and most users require information on displayed persons. The agreement of users on the required anchors is too low (often below 30%) for a ground truth-based evaluation, which is the standard IR evaluation method. As an alternative, we propose a novel evaluation method based on improved search performance and user experience.

KW - METIS-286282

KW - EWI-21603

KW - IR-80184

M3 - Conference contribution

SN - not assigned

SP - 1

EP - 4

BT - Proceedings of the 12th Dutch Belgian Information Retrieval Workshop (DIR 2012)

PB - Ghent University

CY - Ghent

ER -

Aly R, McGuinness K, Kleppe M, Ordelman RJF, O'Connor N, de Jong FMG. Link Anchors in Images: Is there Truth? In Proceedings of the 12th Dutch Belgian Information Retrieval Workshop (DIR 2012). Ghent: Ghent University. 2012. p. 1-4