Research onDistributed Management Method of Motion Picture with Anonymity

2019 Fiscal Year Final Research Report

• Project/Area Number: 17K01149
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Educational technology
• Research Institution: Waseda University
• Principal Investigator: Koyanagi Keiichi 早稲田大学, 理工学術院(情報生産システム研究科・センター), 教授 (20367171)
• Co-Investigator(Kenkyū-buntansha): 土屋 健 公立諏訪東京理科大学, 工学部, 准教授 (90546251) ; 澤野 弘明 愛知工業大学, 情報科学部, 准教授 (10609431)
• Project Period (FY): 2017-04-01 – 2020-03-31
• Keywords: 手話学習 / 分散データ

Outline of Final Research Achievements

In the present study, we studied a feature extraction method for subdividing human movement patterns, and proposed an anonymized distributed information management system that extends the fog computing model. In this method, we focused on the notion of phonemes, which is more detailed than sign language word movements. In this study, we developed a recognition method based on the three elements of the handprint, position, and motion. In anonymized distributed information management, each user's access privileges and the characteristics contained in the information are quantified. It can controll access and building a training model with a small number of distributed data and anonymizing it. We developed a mechanism to provide a combined learning one.

Free Research Field

分散協調システム

Academic Significance and Societal Importance of the Research Achievements

本研究では，匿名性を持った動作特徴抽出モデルを，それぞれのノードで学習して，複数のノードで共有・融合させる仕組みを開発した．この仕組みでは，様々な撮影状況で得られた学習モデルを共有できるため，新たに環境を用意して学習する必要がなくなるため，活用できる範囲を広げるという意味で社会的意義が高い．また，これらの開発の要素技術は，論文として報告しており，学術的意義があるといえる．