Computer vision based recognition of interactions between human body and object

Masumi Kobana*, Jun Ohya

*Corresponding author for this work

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

Abstract

This paper proposes a computer vision based method that recognizes interactions between human body and object. In two successive frames in a video sequence, our function based contour analysis method checks whether the silhouettes of the human body and object overlap. If they do not overlap, our method judges that the human body and object does not interact. If the two silhouettes overlap, our method checks whether the motion vectors obtained from the areas for the human body and object coincide. If they coincide, our method judges that the human body and object are interacting. Some experimental results show the effectiveness of the proposed method.

Original languageEnglish
Title of host publicationEntertainment Computing
Subtitle of host publicationTechnologies and Applications - IFlP 1st International Workshop on Entertainment Computing, IWEC 2002
PublisherSpringer New York LLC
Pages173-180
Number of pages8
ISBN (Print)9781475751536
Publication statusPublished - 2003 Jan 1
EventIFIP 1st International Workshop on Entertainment Computing, IWEC 2002 - Makuhari, Japan
Duration: 2002 May 142002 May 17

Publication series

NameIFIP Advances in Information and Communication Technology
Volume112
ISSN (Print)1868-4238

Conference

ConferenceIFIP 1st International Workshop on Entertainment Computing, IWEC 2002
Country/TerritoryJapan
CityMakuhari
Period02/5/1402/5/17

Keywords

  • Computer vision
  • Motion vector
  • Recognition of interactions
  • Silhouette analysis

ASJC Scopus subject areas

  • Information Systems
  • Computer Networks and Communications
  • Information Systems and Management

Fingerprint

Dive into the research topics of 'Computer vision based recognition of interactions between human body and object'. Together they form a unique fingerprint.

Cite this