DCT-based adaptive metric learning model using asymptotic local information measure

Takami Satonaka, Takaaki Baba, Takayuki Chikamura, Tatsuo Otsuki, Teresa H. Meng

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

1 Citation (Scopus)

Abstract

We present an adaptive metric learning vector quantization procedure based on the discrete-cosine transform (DCT) for accurate face recognition used in multimedia application. Since the set of learning samples may be small, we employ a mixture model of prior distributions. The model selection method, which minimizes the cross entropy between the real distribution and the modeled one, is presented to optimize the mixture number and local metric parameters. The structural risk minimization is used to facilitate an asymptotic approximation of the cross entropy for models of fixed complexity. We also provide a formula to estimate the model complexity derived from the minimum description length criterion. The structural risk minimization method proposed achieves an recognition error rate of 2.29% using the ORL database, which is better than previously reported numbers using the Karhunen-Loeve transform convolution network, the hidden Marcov model and the eigenface model.

Original languageEnglish
Title of host publicationNeural Networks for Signal Processing - Proceedings of the IEEE Workshop
Place of PublicationPiscataway, NJ, United States
PublisherIEEE
Pages521-530
Number of pages10
Publication statusPublished - 1997
Externally publishedYes
EventProceedings of the 1997 7th IEEE Workshop on Neural Networks for Signal Processing, NNSP'97 - Amelia Island, FL, USA
Duration: 1997 Sep 241997 Sep 26

Other

OtherProceedings of the 1997 7th IEEE Workshop on Neural Networks for Signal Processing, NNSP'97
CityAmelia Island, FL, USA
Period97/9/2497/9/26

ASJC Scopus subject areas

  • Signal Processing
  • Software
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'DCT-based adaptive metric learning model using asymptotic local information measure'. Together they form a unique fingerprint.

Cite this