Classification via Incoherent Subspaces

TL;DR

This paper introduces a novel classification method based on incoherent subspaces, enabling class-specific feature extraction and demonstrating competitive performance and efficiency on standard face recognition datasets.

Contribution

The paper proposes a new classification framework using incoherent subspaces with an algorithm for their extraction, advancing feature-based classification techniques.

Findings

Competitive classification accuracy on AR and YaleB datasets

Faster than existing methods like LDA and $\, ext{l}_1$ minimisation

Theoretical analysis of incoherent subspace models

Abstract

This article presents a new classification framework that can extract individual features per class. The scheme is based on a model of incoherent subspaces, each one associated to one class, and a model on how the elements in a class are represented in this subspace. After the theoretical analysis an alternate projection algorithm to find such a collection is developed. The classification performance and speed of the proposed method is tested on the AR and YaleB databases and compared to that of Fisher's LDA and a recent approach based on on $\ell_1$ minimisation. Finally connections of the presented scheme to already existing work are discussed and possible ways of extensions are pointed out.