Fall detection based on skeleton extraction

Zhen Peng Bian, Lap Pui Chau, Nadia Magnenat-Thalmann

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

20 Citations (Scopus)

Abstract

This paper presents an improved skeleton extraction from depth video for fall detection based on fast randomized decision forest (RDF) algorithm. Due to the human's body orientation changes dramatically during falling, it reduces the accuracy of tracking. The human's orientation needs to be corrected before the process by RDF. A rotation to correct the orientation is required frame by frame. Experimental results show that with the help of correction our proposed fall detection method could outperform the existing RDF based method.

Original languageEnglish
Title of host publicationProceedings - VRCAI 2012
Subtitle of host publication11th ACM SIGGRAPH International Conference on Virtual-Reality Continuum and Its Applications in Industry
Pages91-94
Number of pages4
DOIs
Publication statusPublished - Dec 2012
Externally publishedYes
Event11th ACM SIGGRAPH International Conference on Virtual-Reality Continuum and Its Applications in Industry, VRCAI 2012 - Singapore, Singapore
Duration: 2 Dec 20124 Dec 2012

Publication series

NameProceedings - VRCAI 2012: 11th ACM SIGGRAPH International Conference on Virtual-Reality Continuum and Its Applications in Industry

Conference

Conference11th ACM SIGGRAPH International Conference on Virtual-Reality Continuum and Its Applications in Industry, VRCAI 2012
Country/TerritorySingapore
CitySingapore
Period2/12/124/12/12

Keywords

  • 3D
  • activity recognition
  • computer vision
  • fall detection
  • motion analysis
  • motion capture
  • video surveillance

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications

Cite this