Premium
Generating hand motion from body motion based on hand pose estimation
Author(s) -
Oshita Masaki,
Senju Yuta
Publication year - 2016
Publication title -
computer animation and virtual worlds
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.225
H-Index - 49
eISSN - 1546-427X
pISSN - 1546-4261
DOI - 10.1002/cav.1730
Subject(s) - computer science , motion (physics) , artificial intelligence , pose , computer vision , animation , key (lock) , articulated body pose estimation , motion capture , motion estimation , line (geometry) , computer animation , range (aeronautics) , character animation , 3d pose estimation , computer graphics (images) , mathematics , geometry , materials science , computer security , composite material
In this paper, we propose a method to generate hand motion from full‐body motion. We assume that hand pose can be estimated from full‐body pose. We use a support vector machine to choose one of four key hand poses based on the full‐body pose. The training model is constructed from a number of full‐body motions with manually specified hand‐pose keyframes. The output poses are synthesized to include the generated continuous hand motion. We realize smooth transitions between key hand poses and add small movements of fingers for generating plausible hand motions. Because our method does not require future poses, it is valid for both on‐line and off‐line animation. Our experimental results show that our model is applicable to a wide range of full‐body motions.