Keyframe Extraction Using Linear Rotation Invariant Coordinates

Authors : Hasan MUTLU, Ufuk ÇELİKCAN

Pages : 1040-1051

Doi:10.16984/saufenbilder.1148511

View : 21 | Download : 9

Publication Date : 2022-10-20

Article Type : Research Paper

Abstract :Keyframe extraction is a widely applied remedy for issues faced with 3D motion capture -based computer animation. In this paper, we propose a novel keyframe extraction method, where the motion is represented in linear rotation invariant coordinates and the dimensions covering 95% of the data are automatically selected using principal component analysis. Then, by K-means classification, the summarized data is clustered and a keyframe is extracted from each cluster based on cosine similarity. To validate the method, an online user study was conducted. The results of the user study show that 45% of the participants preferred the keyframes extracted using the proposed method, outperforming the alternative by 6%.
Keywords : Keyframe extraction, linear rotation invariant coordinates, motion data summarization