Astronomical high-contrast imaging of circumstellar disks: MUSTARD inverse-problem versus PCA-based methods

TL;DR

This paper introduces the MUSTARD inverse-problem algorithm for high-contrast imaging of circumstellar disks and compares its performance to PCA-based methods, highlighting advantages in recovering disk features.

Contribution

The paper presents the new MUSTARD inverse-problem algorithm and systematically compares its effectiveness against PCA-based methods for disk imaging.

Findings

MUSTARD outperforms PCA-based methods in recovering extended disk features.

Inverse-problem approaches show advantages over PCA in image processing.

Systematic tests demonstrate the robustness of MUSTARD in various observational conditions.

Abstract

Recent observations have shown that protoplanetary disks around young stars can embed a wide variety of features. Raw disk images produced by high-contrast imaging instruments are corrupted by slowly varying residual stellar light in the form of quasi-static speckles. Hence, image processing is required to remove speckles from images and to recover circumstellar signals. Current algorithms that rely on the mainstream angular differential imaging (ADI) observing technique are however limited by geometrical biases, and therefore face a major challenge to reliably infer the morphology of extended disk features. In the last two years, four algorithms have been developed for this task, with three of them based on inverse problem (IP) approaches: REXPACO, MAYONNAISE, and \MUSTARD. In this presentation, we will (i) present the new MUSTARD algorithm and (ii) discuss the advantages of IP compared to others methods based on systematic tests.