Feature selection under learning to rank model for multimedia retrieve

Changsheng Li*, Ling Shao, Changsheng Xu, Hanqing Lu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Citations (Scopus)

Abstract

Most multimedia retrieval problem can be described by ranking model, i.e. the images in the database could be ranked according to the similarity compared with the query image. Existing ranking models generally use the features that are pre-defined by experts. This paper utilized machine learning techniques to automatically select useful features for ranking. We first generate a set of feature subsets by putting each feature into an individual feature subset. Then we sort these feature subsets according to the ranking performances. Third, two neighbor feature subsets in the ranked order are pairwised to generate a new feature subset. The new feature subsets are sorted based on the new ranking performance. Iterate until reach the pre-defined stop point. Experimental results on .gov dataset and Caltech101 development set show the effectiveness and efficiency of the proposed algorithm.

Original languageEnglish
Title of host publicationProceedings of the 2nd International Conference on Internet Multimedia Computing and Service, ICIMCS'10
Pages69-72
Number of pages4
DOIs
Publication statusPublished - 2010
Externally publishedYes
Event2nd International Conference on Internet Multimedia Computing and Service, ICIMCS 2010 - Harbin, China
Duration: 30 Dec 201031 Dec 2010

Publication series

NameProceedings of the 2nd International Conference on Internet Multimedia Computing and Service, ICIMCS'10

Conference

Conference2nd International Conference on Internet Multimedia Computing and Service, ICIMCS 2010
Country/TerritoryChina
CityHarbin
Period30/12/1031/12/10

Keywords

  • Evaluation measure
  • Feature selection
  • Learning to rank
  • Multimedia retrieval

Fingerprint

Dive into the research topics of 'Feature selection under learning to rank model for multimedia retrieve'. Together they form a unique fingerprint.

Cite this