On Geodesic Exponential Kernels

Aasa Feragen, François Lauze, Søren Hauberg

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

Abstract

This extended abstract summarizes work presented at CVPR 2015 [1].

Standard statistics and machine learning tools require input data residing in a Euclidean space. However, many types of data are more faithfully represented in general nonlinear metric spaces or Riemannian manifolds, e.g. shapes, symmetric positive definite matrices, human poses or graphs. The underlying metric space captures domain specific knowledge, e.g. non-linear constraints, which is available a priori. The intrinsic geodesic metric encodes this knowledge, often leading to improved statistical models.
Original languageEnglish
Title of host publicationProceedings of the 3rd International Workshop on Similarity-Based Pattern Recognition (SIMBAD 2015)
EditorsAasa Feragen, Marco Loog, Marcello Pelillo
PublisherSpringer
Publication date2015
Pages211-213
ISBN (Print)978-3-319-24260-6
ISBN (Electronic)978-3-319-24261-3
DOIs
Publication statusPublished - 2015
Event3rd International Workshop on Similarity-Based Pattern Recognition (SIMBAD 2015) - Copenhagen, Denmark
Duration: 12 Oct 201514 Oct 2015
Conference number: 3
http://www.dsi.unive.it/~simbad/2015/

Workshop

Workshop3rd International Workshop on Similarity-Based Pattern Recognition (SIMBAD 2015)
Number3
CountryDenmark
CityCopenhagen
Period12/10/201514/10/2015
Internet address
SeriesLecture Notes in Computer Science
Volume9370
ISSN0302-9743

Fingerprint Dive into the research topics of 'On Geodesic Exponential Kernels'. Together they form a unique fingerprint.

Cite this