Label propagation on data with multiple representations through multi-graph locality preserving projections

Olga Zoidi, Nikos Nikolaidis, Ioannis Pitas

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

135 Downloads (Pure)

Abstract

In this paper a novel method is introduced for propagating label information on data with multiple representations. The method performs dimensionality reduction of the data by calculating a projection matrix that preserves locality
information and a priori pairwise information, in the form of must-link and cannot-link constraints between the various data representations. The final data representations are then fused, in order to perform label propagation. The performance of the proposed method was evaluated on facial images extracted from stereo movies and on the UCF11 action recognition database. Experimental results showed that the proposed method outperforms state of the art methods.
Original languageEnglish
Title of host publication2014 IEEE International Conference on Image Processing (ICIP 2014)
Subtitle of host publicationProceedings of a meeting held 27-30 October 2014, Paris, France
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1505-1509
Number of pages5
ISBN (Electronic)9781479957514
ISBN (Print)9781479957521
Publication statusPublished - Mar 2015
EventIEEE International Conference on Image Processing (ICIP) - Paris, France
Duration: 27 Oct 201430 Oct 2014

Publication series

NameIEEE International Conference on Image Processing (ICIP)
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
ISSN (Print)1522-4880

Conference

ConferenceIEEE International Conference on Image Processing (ICIP)
CountryFrance
CityParis
Period27/10/1430/10/14

Keywords

  • Locality preserving projections
  • dimensionality reduction
  • label propagation
  • multiple graphs

Fingerprint Dive into the research topics of 'Label propagation on data with multiple representations through multi-graph locality preserving projections'. Together they form a unique fingerprint.

Cite this