Paper 14145-99
Paper 14145-99
Coherence differential imaging using random forest regression for the direct detection of exoplanets
10 July 2026 • 14:50 - 15:10 CEST | Room B4-M3
Abstract
Coronagraphic imaging of exoplanets is limited by residual speckles that mimic planets. Advanced post-processing is essential for current and future instruments on the ground or in space.
Current techniques are time-intensive and limited. ADI requires long sequences and is limited at small separation. RDI is also time-consuming and sensitive to speckle evolution, leading to imperfect subtraction. Coherence Differential Imaging (CDI) offers a faster alternative by using the light incoherence between speckles and planets, but its reliance on accurate instrumental models limits performance.
After demonstration on SPHERE, we enhanced CDI with a Random Forest algorithm. This method models the differential signal as a high-dimensional regression problem, optimizing the discrimination between coherent speckle noise and the incoherent planet signal. We validate this enhanced CDI using testbed data reproducing Roman CGI conditions.
Presenter
Johan Mazoyer
Observatoire de Paris (France)
Johan Mazoyer is a tenured CNRS Research Scientist at Paris Observatory / PSL, since 2020. He earned his PhD in 2014 from Université Paris Cité. His research lies at the intersection of optical instrumentation and observational astrophysics. He develops and implements advanced high-contrast imaging and coronagraphy techniques — including active wavefront control — with the aim of directly detecting and characterizing exoplanets and debris disks. In parallel, he analyses coronagraphic and high-contrast images from ground-based and space instruments to study debris disks (analogs of solar system Kuiper belts), shedding light on the late stages of planetary system formation and evolution. He is the deputy French representative in the Roman coronagraphic instrument community participation program. He was recently awarded an ERC consolidator project to push forward the limit of future space and ground-based coronagraphic instruments.