Segmentation of Subspaces in Sequential Data

TL;DR

This paper introduces Ordered Subspace Clustering (OSC), a novel method for segmenting sequential data into subspaces, outperforming existing techniques across various data types like hyperspectral, video, and motion capture.

Contribution

The paper presents OSC, a new clustering approach that incorporates sequential information into sparse subspace segmentation, improving accuracy over prior methods.

Findings

OSC outperforms SpatSC, LRR, and SSC in experiments.

Effective on hyperspectral, video, and motion capture data.

Incorporates sequential order into subspace clustering.

Abstract

We propose Ordered Subspace Clustering (OSC) to segment data drawn from a sequentially ordered union of subspaces. Similar to Sparse Subspace Clustering (SSC) we formulate the problem as one of finding a sparse representation but include an additional penalty term to take care of sequential data. We test our method on data drawn from infrared hyper spectral, video and motion capture data. Experiments show that our method, OSC, outperforms the state of the art methods: Spatial Subspace Clustering (SpatSC), Low-Rank Representation (LRR) and SSC.