Unsupervised Dataset Dictionary Learning for domain shift robust clustering: application to sitting posture identification

TL;DR

This paper proposes U-DaDiL, an unsupervised dataset dictionary learning method that improves clustering robustness across domains, specifically applied to sitting posture identification, by aligning dataset distributions using Wasserstein barycenters.

Contribution

Introduces U-DaDiL, a novel unsupervised clustering approach that effectively handles domain shifts through distribution alignment with Wasserstein barycenters.

Findings

Significant improvement in cluster alignment accuracy on Office31 dataset

Effective handling of domain shift in unsupervised sitting posture identification

Demonstrates robustness of the method across diverse datasets

Abstract

This paper introduces a novel approach, Unsupervised Dataset Dictionary Learning (U-DaDiL), for totally unsupervised robust clustering applied to sitting posture identification. Traditional methods often lack adaptability to diverse datasets and suffer from domain shift issues. U-DaDiL addresses these challenges by aligning distributions from different datasets using Wasserstein barycenter based representation. Experimental evaluations on the Office31 dataset demonstrate significant improvements in cluster alignment accuracy. This work also presents a promising step for addressing domain shift and robust clustering for unsupervised sitting posture identification