Online sparse matrix gaussian process regression and vision applications

Ananth Ranganathan, Ming Hsuan Yang

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

12 Citations (Scopus)

Abstract

We present a new Gaussian Process inference algorithm, called Online Sparse Matrix Gaussian Processes (OSMGP), and demonstrate its merits with a few vision applications. The OSMGP is based on the observation that for kernels with local support, the Gram matrix is typically sparse. Maintaining and updating the sparse Cholesky factor of the Gram matrix can be done efficiently using Givens rotations. This leads to an exact, online algorithm whose update time scales linearly with the size of the Gram matrix. Further, if approximate updates are permissible, the Cholesky factor can be maintained at a constant size using hyperbolic rotations to remove certain rows and columns corresponding to discarded training examples. We demonstrate that, using these matrix downdates, online hyperparameter estimation can be included without affecting the linear runtime complexity of the algorithm. The OSMGP algorithm is applied to head-pose estimation and visual tracking problems. Experimental results demonstrate that the proposed method is accurate, efficient and generalizes well using online learning.

Original languageEnglish
Title of host publicationComputer Vision - ECCV 2008 - 10th European Conference on Computer Vision, Proceedings
Pages468-482
Number of pages15
EditionPART 1
DOIs
Publication statusPublished - 2008 Dec 1
Event10th European Conference on Computer Vision, ECCV 2008 - Marseille, France
Duration: 2008 Oct 122008 Oct 18

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume5302 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th European Conference on Computer Vision, ECCV 2008
CountryFrance
CityMarseille
Period08/10/1208/10/18

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Online sparse matrix gaussian process regression and vision applications'. Together they form a unique fingerprint.

  • Cite this

    Ranganathan, A., & Yang, M. H. (2008). Online sparse matrix gaussian process regression and vision applications. In Computer Vision - ECCV 2008 - 10th European Conference on Computer Vision, Proceedings (PART 1 ed., pp. 468-482). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5302 LNCS, No. PART 1). https://doi.org/10.1007/978-3-540-88682-2-36