Non-linear weighted averaging for multimodal information fusion by employing Analytical Network Process

Turgay Yilmaz, Adnan Yazici, Masaru Kitsuregawa

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

3 Citations (Scopus)

Abstract

Linear combination is a popular approach in information fusion due to its simplicity. However, it suffers from the performance upper-bound of linearity and dependency on the selection of weights. In this study, we introduce a 'simple' alternative for linear combination, which is a non-linear extension on it. The approach is based on the Analytical Network Process, which is a popular approach in Operational Research, but never applied for fusion before. The approach benefits from two major ideas; interdependency between classes and dependency of classes on the features. Experiments conducted on CCV dataset demonstrate that proposed approach outperforms linear combination and other simple approaches, moreover it is less-dependent on the selection of weights.

Original languageEnglish
Title of host publicationICPR 2012 - 21st International Conference on Pattern Recognition
Pages234-237
Number of pages4
Publication statusPublished - Dec 1 2012
Event21st International Conference on Pattern Recognition, ICPR 2012 - Tsukuba, Japan
Duration: Nov 11 2012Nov 15 2012

Publication series

NameProceedings - International Conference on Pattern Recognition
ISSN (Print)1051-4651

Conference

Conference21st International Conference on Pattern Recognition, ICPR 2012
CountryJapan
CityTsukuba
Period11/11/1211/15/12

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Fingerprint Dive into the research topics of 'Non-linear weighted averaging for multimodal information fusion by employing Analytical Network Process'. Together they form a unique fingerprint.

Cite this