3D Interest Point Detection using Local Surface Characteristics with Application in Action Recognition

Michael Boelstoft Holte

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

5 Citations (Scopus)

Abstract

In this paper we address the problem of detecting 3D inter- est points (IPs) using local surface characteristics. We con- tribute to this field by introducing a novel approach for detec- tion of 3D IPs directly on a surface mesh without any require- ments of additional image/video information. The proposed Difference-of-Normals (DoN) 3D IP detector operates on the surface mesh, and evaluates the surface structure (curvature) locally (per vertex) in the mesh data. We present an exam- ple of application in action recognition from a sequence of 3-dimensional geometrical data, where local 3D motion de- scriptors, Histogram of Optical 3D Flow (HOF3D), are ex- tracted from estimated 3D optical flow in the neighborhood of each IP and made view-invariant. Experiments on the pub- licly available i3DPost dataset show promising results.
Original languageEnglish
Title of host publicationIEEE International Conference on Image Processing
PublisherIEEE Signal Processing Society
Publication date2014
Pages5736-5740
ISBN (Print)978-1-4799-5751-4
DOIs
Publication statusPublished - 2014
EventIEEE International Conference on Image Processing (ICIP) - Paris, France
Duration: 27 Oct 201430 Oct 2014

Conference

ConferenceIEEE International Conference on Image Processing (ICIP)
Country/TerritoryFrance
CityParis
Period27/10/201430/10/2014

Fingerprint

Dive into the research topics of '3D Interest Point Detection using Local Surface Characteristics with Application in Action Recognition'. Together they form a unique fingerprint.

Cite this