Spectral clustering of linear subspaces for motion segmentation

Spectral clustering of linear subspaces for motion segmentation
复制标题

DOI:
10.1109/iccv.2009.5459173
复制
发表时间:
2009-09
期刊:
2009 IEEE 12th International Conference on Computer Vision
影响因子:
--
通讯作者:
Fabien Lauer;C. Schnörr
Fabien Lauer;C. Schnörr
中科院分区:
其他
文献类型:
--
作者:
Fabien Lauer;C. Schnörr

文献摘要

被引文献

相似文献

本文研究了通过谱嵌入和线性子空间聚类的方法从跟踪的特征点中自动分割出多个运动。我们证明了环境空间的维度对于可分离性是至关重要的,并且前人工作中选择的低维度并不是最优的。我们建议使用数据驱动的方法来选择最优的环境维度,以确定最优的环境维度。将我们的方法应用于Hopkins155视频基准数据库,在分割精度和计算速度方面都一致地优于一系列最先进的方法。
This paper studies automatic segmentation of multiple motions from tracked feature points through spectral embedding and clustering of linear subspaces. We show that the dimension of the ambient space is crucial for separability, and that low dimensions chosen in prior work are not optimal. We suggest lower and upper bounds together with a data-driven procedure for choosing the optimal ambient dimension. Application of our approach to the Hopkins155 video benchmark database uniformly outperforms a range of state-of-the-art methods both in terms of segmentation accuracy and computational speed.