Extraction and visualization of the important features of motions by experts for supporting training of motion forms.

2019 Fiscal Year Final Research Report

• Project/Area Number: 15H02704
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Multimedia database
• Research Institution: Kyushu Institute of Technology
• Principal Investigator: Oshita Masaki 九州工業大学, 大学院情報工学研究院, 准教授 (20363400)
• Co-Investigator(Kenkyū-buntansha): 向井 智彦 首都大学東京, システムデザイン研究科, 准教授 (10432296) ; 栗山 繁 豊橋技術科学大学, 工学(系)研究科(研究院), 教授 (20264939)
• Project Period (FY): 2015-04-01 – 2020-03-31
• Keywords: 動作解析・評価 / 機械学習 / 可視化

Outline of Final Research Achievements

The purpose of this research was to extract the important features of motions by experts such as athletes, dancers, and craftsmen in order to utilize them for novices to learn how to move like experts. We had developed various methods for computation of kinematics motions features, selection of important motion features by applying machine learning, construction of statistical models of motion features and their relationships, visualization of spatial ranges of motions, visualization of the difference between the motion model and an input motion, and generation and deformation of motions of experts by several approaches such as deep learning, optimization, and motion interpolation. Based on these methods, we had developed a prototype of motion training for testis shot forms and evaluated the effectiveness of our system through a user study.

Free Research Field

コンピュータアニメーション

Academic Significance and Societal Importance of the Research Achievements

本研究の成果により、熟練者の動作データの特徴を抽出して可視化したり、練習者の動作との差異を可視化したりすることで、初心者が熟練者の動作を効率的に習得することができるシステムを実現できた。感覚的なものであり具体的に伝えることが難しいと言われていた熟練者の動作のコツを、情報工学的なアプローチで扱えることを明らかにできた。また、本研究の成果である、深層学習・最適化問題・動作補間の複数のアプローチによる熟練者の動作の生成・変形手法は、一般的なアニメーション制作にも応用が期待できる。