People Image Analysis by Video Understanding Technology

DOI Web Site 12 References

Bibliographic Information

Other Title
  • 動画像理解による人を観る技術(セキュリティカメラ,車載カメラ,サーベイランス技術,関連デバイス,および一般)
  • 招待講演 動画像理解による人を観る技術
  • ショウタイ コウエン ドウガゾウ リカイ ニ ヨル ヒト オ ミル ギジュツ

Search this article

Abstract

Automatic people detection is a key enabler for applications in visual surveillance. A number of methods have been developed based on background subtraction for detecting motion from images captured by fixed cameras. Because motion detection-based methods such as background subtraction use a top-down approach, object classi-fication in the next step becomes impossible if the object's region is not segmented correctly. A window-scanning approach, which is a combination of low-level features and statistical learning, has been proposed for solving this problem because of the great improvement in computer speed over the last several years. In this paper, we describe the conventional approach of a video understanding algorithm, developed during the VSAM project, for detecting motion, and we introduce applications that use it. We also introduce an example of how a window-scanning approach for detecting people can be implemented as a novel approach for analyzing the images of people.

Journal

  • ITE Technical Report

    ITE Technical Report 32.32 (0), 13-16, 2008

    The Institute of Image Information and Television Engineers

References(12)*help

See more

Keywords

Details 詳細情報について

Report a problem

Back to top