Graph Constrained Data Representation Learning for Human Motion Segmentation

TL;DR

This paper introduces an unsupervised data representation learning model for human motion segmentation that outperforms existing transfer learning methods by learning a flexible auxiliary data matrix with local structure preservation.

Contribution

The novel model learns an auxiliary data matrix and a coding scheme jointly, enhancing clustering without relying on prior domain knowledge, unlike transfer subspace methods.

Findings

Achieves superior clustering performance on four benchmark datasets.

Outperforms state-of-the-art transfer learning approaches.

Efficient optimization via a new ADMM formulation.

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, we learn an auxiliary data matrix that can deviate from the initial data, hence confer more degrees of freedom to the coding matrix. Second, we introduce a regularization term for this auxiliary data matrix 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 auxiliary 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.