Browse DORAS
Browse Theses
Search
Latest Additions
Creative Commons License
Except where otherwise noted, content on this site is licensed for use under a:

Enhancing person annotation for personal photo management applications

Cooray, Saman H. and O'Connor, Noel E. (2009) Enhancing person annotation for personal photo management applications. In: MDMM 2009 - 3rd International Workshop on Multimedia Data Mining and Management, 31 August - 4 September 2009, Linz, Austria. ISBN 978-0-7695-3763-4

Full text available as:

[img]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
1121Kb

Abstract

This paper addresses a sub-problem of the broad annotation problem, namely "person annotation", associated with personal digital photo management and investigates approaches to enhancing person annotation in personal photo management applications. We study a number of approaches to enhance the performance of semi-automatic person annotation using real-life personal photo collections as the test data. To this end, face and body-patch features are employed to describe the appearance of a person as a means to more effectively capture the identities of re-appearing people in personal photo archives. Experiments are carried out to identify a suitable initial annotation method, compare the performances of event-constrained person matching with global person matching, and the effect of the size of initial annotation on the overall performance of person annotation in real-life personal photo archives. The evaluation results, presented in terms of H-Hit rate figures, illustrate that using event-constrained person matching with event-based initial annotation proves to be a better performing approach than global person matching for person annotation in personal photo archives. Results also clearly demonstrate the nature of compromise that needs to be made when annotating large photo collections in terms of accuracy against user-interaction.

Item Type:Conference or Workshop Item (Paper)
Event Type:Workshop
Refereed:Yes
Uncontrolled Keywords:personal photo management; person annotation; face recognition; content-based descriptors;
Subjects:Computer Science > Multimedia systems
DCU Faculties and Centres:DCU Faculties and Schools > Faculty of Engineering and Computing > School of Electronic Engineering
Research Initiatives and Centres > CLARITY: The Centre for Sensor Web Technologies
Official URL:http://dx.doi.org/10.1109/DEXA.2009.41
Copyright Information:©2009 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Funders:Science Foundation Ireland
ID Code:14799
Deposited On:11 Aug 2009 09:53 by Saman Cooray. Last Modified 29 Apr 2010 13:35

Download statistics

Archive Staff Only: edit this record