Publication

Graph constrained data representation learning for human motion segmentation

Conference Article

Conference

International Conference on Computer Vision (ICCV)

Edition

2021

Pages

1440-1449

Doc link

https://doi.org/10.1109/ICCV48922.2021.00149

File

Download the digital copy of the doc pdf document

Abstract

Recently, transfer subspace learning based approaches have shown to be a valid alternative to unsupervised subspace clustering and temporal data clustering for human motion segmentation (HMS). These approaches leverage prior knowledge from a source domain to improve clustering performance on a target domain, and currently they represent the state of the art in HMS. Bucking this trend, in this paper, we propose a novel unsupervised model that learns a representation of the data and digs clustering information from the data itself. Our model is reminiscent of temporal subspace clustering, but presents two critical differences. First, it confers to the coding matrix more degree of freedom by allowing the data representation matrix to deviate from the initial data. Second, it introduces a regularization term on the data representation, that preserves the local geometrical structure present in the high-dimensional space. The proposed model is efficiently optimized by using an original Alternating Direction Method of Multipliers (ADMM) formulation allowing to learn jointly the data representation, a nonnegative dictionary and a coding matrix. Experimental results on four benchmark datasets for HMS demonstrate that our approach achieves significantly better clustering performance then state-of-the-art methods, including both unsupervised and more recent semi-supervised transfer learning approaches.

Categories

pattern recognition.

Author keywords

graph based regularization, temporal subspace clustering, human motion segmentation

Scientific reference

M. Dimiccoli, L. Garrido, G. Rodríguez and H. Wendt. Graph constrained data representation learning for human motion segmentation, 2021 International Conference on Computer Vision, 2021, Montreal, Canada, pp. 1440-1449.