Skip to Main content Skip to Navigation
Journal articles

Joint deconvolution and unsupervised source separation for data on the sphere

Abstract : Tackling unsupervised source separation jointly with an additional inverse problem such as deconvolution is central for the analysis of multi-wavelength data. This becomes highly challenging when applied to large data sampled on the sphere such as those provided by wide-field observations in astrophysics, whose analysis requires the design of dedicated robust and yet effective algorithms. We therefore investigate a new joint deconvolution/sparse blind source separation method dedicated for data sampled on the sphere, coined SDecGMCA. It is based on a projected alternate least-squares minimization scheme, whose accuracy is proved to strongly rely on some regularization scheme in the present joint deconvolution/blind source separation setting. To this end, a regularization strategy is introduced that allows designing a new robust and effective algorithm, which is key to analyze large spherical data. Numerical experiments are carried out on toy examples and realistic astronomical data.
Document type :
Journal articles
Complete list of metadata
Contributor : Jerome Bobin <>
Submitted on : Tuesday, March 23, 2021 - 2:41:00 PM
Last modification on : Friday, March 26, 2021 - 3:12:12 AM


Files produced by the author(s)



R. Carloni Gertosio, Jerome Bobin. Joint deconvolution and unsupervised source separation for data on the sphere. Digital Signal Processing, Elsevier, 2021, 110, pp.102946. ⟨10.1016/j.dsp.2020.102946⟩. ⟨hal-03177873⟩



Record views


Files downloads