2022 Fiscal Year Final Research Report
Development of hand motion analysis system for cloth model manipulation using actual finger motion in virtual draping
Project/Area Number |
18K13041
|
Research Category |
Grant-in-Aid for Early-Career Scientists
|
Allocation Type | Multi-year Fund |
Review Section |
Basic Section 08030:Family and consumer sciences, and culture and living-related
|
Research Institution | Nagano National College of Technology |
Principal Investigator |
Mesuda Yuko 長野工業高等専門学校, 情報エレクトロニクス系, 准教授 (20757893)
|
Project Period (FY) |
2018-04-01 – 2023-03-31
|
Keywords | 仮想立体裁断 / 手指動作 / 布モデル操作 |
Outline of Final Research Achievements |
In this study, series of movements in draping are input to AI to confirm which feature value are influenced on the accuracy of the AI. Input movements are "grab","release","fasten" and so on. The result suggest time-series data of hand movements are influenced on the accuracy of the AI more than hand form in an instant in cloth manipulation. Bending data of fingers are important feature for hand movements of cloth manipulation as same as three dimensional coordinates of hand and fingers. Hand movements with actual cloth include complex motions, therefore the discrimination accuracy is low when training with the actual cloth handling data.
|
Free Research Field |
衣服シミュレーション
|
Academic Significance and Societal Importance of the Research Achievements |
曲げ率のデータのみでもつまむ,放す,固定するといった動作を高い精度で判別することができたことから,指間の広がりや手の位置情報だけが動作の特徴を有しているわけではないことが示唆された.三次元座標データの次元に比べ曲げ率データの次元の方が大幅に少ないため,曲げ率データをさらに詳しく解析することで少ない特徴量で各動作を表せる可能性がある. この結果は実際の手指の動作を用いた直観的な布モデル操作の実現に繋がり,最終的には着付けや衣服づくりのオンライン教材やリハビリなどにも応用できる技術の発展に貢献できるものであると考えられえる.
|