Title: Joint unmixing and deconvolution for angular and spectral differential imaging

Abstract: Angular and spectral differential imaging is an observational technique used in astronomy to study the close environment of stars. The relative angular motion and spectral scaling between on-axis and off-axis sources are exploited by processing techniques to separate two components: the residual light from the star and the light coming from surrounding objects such as circumstellar disks. This paper describes a method to unmix these two components and deconvolve the image of disks. It is based on a statistical modeling of the residual star light, in particular its spatial and spectral correlations. These correlations have so far been neglected by standard reconstruction techniques. We show on several datasets from the SPHERE instrument at the Very Large Telescope, Chile, that accounting for correlations strongly improves the reconstructions.