Faster and better sparse blind source separation through mini-batch optimization - Archive ouverte HAL Access content directly
Preprints, Working Papers, ... Year :

Faster and better sparse blind source separation through mini-batch optimization

, , (1)
1
C Kervazo
T Liaudat
  • Function : Author
Jerome Bobin
  • Function : Author
  • PersonId : 858908

Abstract

Sparse Blind Source Separation (sBSS) plays a key role in scientific domains as different as biomedical imaging, remote sensing or astrophysics, which require the development of increasingly faster and scalable BSS methods without sacrificing the separation performances. To that end, a new distributed sparse BSS algorithm is introduced based on a mini-batch extension of the Generalized Morphological Component Analysis algorithm (GMCA). Precisely, it combines a robust projected alternate least-squares method with mini-batches optimization. The originality further lies in the use of a manifold-based aggregation of asynchronously estimated mixing matrices. Numerical experiments are carried out on realistic spectroscopic spectra , and highlight the ability of the proposed distributed GMCA (dGMCA) to provide very good separation results even when very small mini-batches are used. Quite unexpectedly, it can further outperform the (non-distributed) state-of-the-art methods for highly sparse sources.
Fichier principal
Vignette du fichier
distributed_GMCA.pdf (1.71 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-02426991 , version 1 (03-01-2020)

Identifiers

  • HAL Id : hal-02426991 , version 1

Cite

C Kervazo, T Liaudat, Jerome Bobin. Faster and better sparse blind source separation through mini-batch optimization. 2020. ⟨hal-02426991⟩
190 View
151 Download

Share

Gmail Facebook Twitter LinkedIn More