A RELIEF-based modality weighting approach for multimodal information retrieval

Turgay Yilmaz, Elvan Gulen, Adnan Yazici, Masaru Kitsuregawa

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

2 Citations (Scopus)

Abstract

Despite the extensive number of studies for multimodal information fusion, the issue of determining the optimal modalities has not been adequately addressed yet. In this study, a RELIEF-based multimodal feature selection approach (RELIEF-RDR) is proposed. The original RELIEF algorithm is extended for weaknesses in three major issues; multilabeled data, noise and class-specific feature selection. To overcome these weaknesses, discrimination based weighting mechanism of RELIEF is supported with two additional concepts; representation and reliability capabilities of features, without an increase in computational complexity. These capabilities of features are exploited by using the statistics on dissimilarities of training instances. The experiments conducted on TRECVID 2007 dataset validated the superiority of RELIEF-RDR over RELIEF.

Original languageEnglish
Title of host publicationProceedings of the 2nd ACM International Conference on Multimedia Retrieval, ICMR 2012
DOIs
Publication statusPublished - 2012
Externally publishedYes
Event2nd ACM International Conference on Multimedia Retrieval, ICMR 2012 - Hong Kong, China
Duration: Jun 5 2012Jun 8 2012

Publication series

NameProceedings of the 2nd ACM International Conference on Multimedia Retrieval, ICMR 2012

Conference

Conference2nd ACM International Conference on Multimedia Retrieval, ICMR 2012
Country/TerritoryChina
CityHong Kong
Period6/5/126/8/12

Keywords

  • Feature weighting
  • Multimodal information fusion
  • RELIEF

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Vision and Pattern Recognition
  • Software

Fingerprint

Dive into the research topics of 'A RELIEF-based modality weighting approach for multimodal information retrieval'. Together they form a unique fingerprint.

Cite this