• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to project page

2021 Fiscal Year Final Research Report

Person tracking in a wide area via spatio-temporal data dropout

Research Project

  • PDF
Project/Area Number 18K18070
Research Category

Grant-in-Aid for Early-Career Scientists

Allocation TypeMulti-year Fund
Review Section Basic Section 61010:Perceptual information processing-related
Research InstitutionInstitute of Physical and Chemical Research (2021)
Nagoya University (2018-2020)

Principal Investigator

Yasutomo Kawanishi  国立研究開発法人理化学研究所, 情報統合本部, チームリーダー (50755147)

Project Period (FY) 2018-04-01 – 2022-03-31
Keywords人物追跡 / 時空間データドロップアウト / 広域監視 / 人物照合 / アンサンブル学習
Outline of Final Research Achievements

In this research, we developed a method for accurately tracking persons within and across camera views in a scene where a large number of fixed cameras are observing a wide area. We propose a new concept in tracking, "spatio-temporal data dropout," and integrate a large number of tracking results obtained by this method by considering them as "weak tracking results" based on the analogy with ensemble learning. This proposed method for tracking a person over a wide area addresses a critical problem in tracking a person within/camera views, which is that the person is temporarily obscured by changes in illumination or other objects, causing the tracking to be interrupted.

Free Research Field

画像認識

Academic Significance and Societal Importance of the Research Achievements

今や監視カメラはあらゆるところに設置されつつある。監視カメラ映像中の人物がどこから来てどこへ行ったのかを集計するためには、人物を多数のカメラ視野間にわたって追跡する必要があるが、精度良く人物の移動軌跡を得ることは困難である。本研究では、一時的な見えの変化による追跡の失敗を改善することができる手法を提案した。本研究が取り組んだ広域の人物追跡は、防犯やマーケティングなど、様々な分野での利用価値がある。

URL: 

Published: 2023-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi