Improved low rank representation : kernelization, efficient optimization and applications

Given data sampled from multiple subspaces, the goal of subspace clustering is to partition data into several clusters, so that each cluster exactly corresponds to one subspace. Initially proposed for subspace clustering, the low rank representation (LRR) approach has shown promising results in vari...

ver descrição completa

Detalhes bibliográficos
Autor principal: Xiao, Shijie
Outros Autores: Cai Jianfei
Formato: Tese
Idioma:English
Publicado em: 2016
Assuntos:
Acesso em linha:http://hdl.handle.net/10356/66234