Optimized keyframe extraction for 3D character animations

Abstract

In this paper, we propose a new method to automatically extract keyframes from animation sequences. Our method can be applied equally to both skeletal and mesh animations. It uses animation saliency computed on the original data to help select the group of keyframes that can reconstruct the input animation with less perception error. For computational efficiency, we perform nonlinear dimension reduction using locally linear embedding and then carry out the optimal search in much lower‐dimensional space. With this approach, reconstruction of the animation from the extracted keyframes shows much better results as compared with earlier approaches. Copyright © 2012 John Wiley & Sons, Ltd.

Towards key-frame extraction methods for 3D video: a review

“…Most 3D key-frame extraction methods cited in this paper were developed for specific content and 3D format and only four of them include comparisons with similar methods [18,22,41,51]. In [18], curve simplification, UnS and clustering methods were utilised as reference methods for performance evaluation and comparison of the proposed matrix factorisation methods.…”

“…In [18], curve simplification, UnS and clustering methods were utilised as reference methods for performance evaluation and comparison of the proposed matrix factorisation methods. The authors showed that the method based on matrix factorisation extracts more representative key-frames in comparison with the other three competing methods [22,41,51]. However, the algorithm is very slow with quadratic running time complexity.…”

“…Jin et al in [41] compare the proposed method with the UnS and Principal Component Analysis methods [56]. The results show that the proposed method achieves much better reconstruction of skeletal and mesh animation than the other methods under analysis.…”

“…In some summarisation applications, one key-frame per video shot is not enough, and the PoS method can be adapted to this need by allowing the selection of multiple frames at fixed positions within the video shot. For 3D video, UnS and PoS are used mostly as references for comparisons with other methods, as in [18,41,42]. Ionescu et al [36] selected as key-frames the frames in the middle of the video shot to reduce temporal redundancy and computation cost of his animation movies summarisation method.…”

“…Jin et al [41] proposed a key-frame extraction method for animation sequences (skeletal and mesh animations). The method uses animation saliency computed from the original data to aid the selection of the key-frames that can be used to reconstruct the original animation with smaller error.…”