A Possible Converter to Denoise the Images of Exoplanet Candidates through Machine Learning Techniques

It was particularly exciting to have directly imaged exoplanets for the first time in 2008 (Kalas et al., 2008) as it gave signals directly from these exoplanets and thus confirmed their existence. Since then, many research groups have spent considerable efforts to improve the techniques of high-contrast imaging in order to detect more exoplanets (Tamura, 2009; Enya & Abe, 2010; Kuzuhara et al., 2013; Dou et al., 2015; Dou & Ren, 2016). In addition, new high-contrast imaging instruments were developed for eight-meter class telescopes such as the Gemini Planet Imager (GPI) (Macintosh et al., 2006) for Gemini South, the Subaru Coronagraphic Extreme Adaptive Optics (SCExAO) (Jovanovic et al., 2015) for Subaru Telescope, and the Spectro-Polarimetic High contrast imager for Exoplanet Research (SPHERE) (Beuzit et al., 2019) for Very Large Telescope (VLT). Moreover, a new camera was designed for SCExAO to further advance the performance of high contrast imaging (Walter et al., 2020). It is notable that these instruments often bring very interesting related results (Mayama et al., 2006; Itoh et al., 2008). To detect exoplanets through the method of direct imaging, the highcontrast imaging employs the technique of angular differential imaging (ADI) (Marois et al., 2006) and produces many frames with different parallactic angles, i.e.