MODEL&CO: Exoplanet Detection in Angular Differential Imaging Using Machine Learning Across Multiple Observations

T. Bodrito, O. Flasseur, J. Mairal, J. Ponce, M. Langlois, and A. M. Lagrange

MNRAS 534, 1569–1596 (2024)
Advance Access publication September 19, 2024
https://doi.org/10.1093/mnras/stae2174
Abstract

Direct imaging of exoplanets is particularly challenging due to the high contrast between the planet and the star luminosities, and their small angular separation. In addition to tailored instrumental facilities implementing adaptive optics and coronagraphy, post-processing methods combining several images recorded in pupil tracking mode are needed to attenuate the nuisances corrupting the signals of interest. Most of these post-processing methods build a model of the noise based on the target observations themselves, resulting in severely limited detection sensitivity at short angular separations due to the lack of angular diversity. To address this issue, we propose building the noise model from an archive of multiple observations by leveraging supervised deep learning techniques. The proposed approach frames the detection problem as a reconstruction task and captures the structure of the noise from two complementary representations of the data. Unlike methods inspired by reference differential imaging, the proposed model is highly nonlinear and does not rely on explicit image-to-image similarity measurements and subtractions.The proposed approach also incorporates statistical modeling of learnable spatial features. This is beneficial for improving both detection sensitivity and robustness against heterogeneous data. We apply the proposed algorithm to several datasets from the VLT/SPHERE instrument and demonstrate a superior precision-recall trade-off compared to the PACO algorithm. Interestingly, the improvement is particularly significant when the diversity induced by ADI is most limited, thus supporting the ability of the proposed approach to learn information across multiple observations.