Modeling image appeal based on crowd preferences for automated person-centric collage creation

Vassilios Vonikakis, Ramanathan Subramanian, Jonas Arnfred, Stefan Winkler

Research output: A Conference proceeding or a Chapter in BookConference contributionpeer-review

4 Citations (Scopus)

Abstract

This paper attempts to model Image Appeal of photos por- traying people, with the objective of automatically rank- ing and selecting the most appealing ones for the creation of interesting person-centric collages/collections. To under- stand the notion of image appeal, we employed crowdsourc- ing, using 350 workers who were asked to select a repre- sentative subset of images from five different person-centric album themes (involving a man, woman, couple, girl and baby). The albums were previously balanced with respect to nine different image attributes using Binary Integer Pro- gramming. The crowdsourcing study revealed identifiable patterns in the photo selection process, with more appeal- ing photos securing more hits than less appealing ones. We then employed nine low-level image features and Support Vector Regressors to model photo selection statistics{ the best model explained 63% of the selection patterns, and our analyses also confirmed the role of context in inuencing Image Appeal. Finally, Image Appeal predictions on unseen photos are presented to demonstrate the promise of our ap- proach.

Original languageEnglish
Title of host publicationCrowdMM 2014 - Proceedings of the International Workshop on Crowdsourcing for Multimedia, Workshop of MM 2014
EditorsJudith Redi, Mathias Lux
Place of PublicationUnited States
PublisherAssociation for Computing Machinery (ACM)
Pages9-15
Number of pages7
ISBN (Electronic)9781450331289
DOIs
Publication statusPublished - 7 Nov 2014
Externally publishedYes
Event3rd International ACM Workshop on Crowdsourcing for Multimedia, CrowdMM 2014 - Orlando, United States
Duration: 7 Nov 2014 → …

Publication series

NameCrowdMM 2014 - Proceedings of the International Workshop on Crowdsourcing for Multimedia, Workshop of MM 2014

Conference

Conference3rd International ACM Workshop on Crowdsourcing for Multimedia, CrowdMM 2014
Country/TerritoryUnited States
CityOrlando
Period7/11/14 → …

Fingerprint

Dive into the research topics of 'Modeling image appeal based on crowd preferences for automated person-centric collage creation'. Together they form a unique fingerprint.

Cite this