Hello Research project Publications Code Communications Sand-jrd github cat

About me


My name is Sandrine Juillard, and I am currently pursuing my PhD at the University of Liège - in The Planetary & Stellar system Imaging Laboratory (PSILab) - under the supervision of Valentin Christiaens and Olivier Absil.

I am conducting research in the field of signal processing applied to high-contrast imaging. I recommend using Iterative Principal Component Analysis (I-PCA), also known as GreeDS, for processing ADI/RDI infrared scattered-light HCI datasets (e.g VLT/SHPERE/IRDIS, KECK/NIRC). Why? see Juillard et. al 2023.
You can install my implementation of GreeDS, which is fast, simple, and requires PyTorch. See the instructions below or in the "code" section.

Currently, in my final year, I started a new project in collaboration with Yannis Argyriou and Danny Gasman (KU Leuven, Belgium), focused on PSF-subtraction for the JWST/MIRI-MRS instrument.

As I approach the completion of my Ph.D. in September 2025, I am actively seeking to pursue a career in space science.

pet picture

Posters


Check out my posters for a quick overview of my research.
Open the poster in a PDF viewer Related paper
open tab A Spiral Arm in the outer disk of PDS 70 ? Juillard et. al 2022
open tab Inverse-problem versus Principal Component Analysis Juillard et. al 2023
open tab Combining reference-star and angular differential imaging Juillard et. al 2024
open tab JWST/MIRI/MRS: PSF Subtraction in the Detector Plane -

Iterative PCA


Download IPCA via PyPI:

pip install GreeDS

Exemple usage

    
Import package and load data
from GreeDS import GreeDS from vip_hci.fits import open_fits cube = open_fits( "your_cube.fits") angles = open_fits( "your_PA_angles.fits") ref = open_fits( "your_refs.fits")
    
Use IPCA
r = 10 # Iteration over PCA-rank l = 10 # Iteration per rank r_start = 1 # PCA-rank to start iteration (good for faint signals) pup_size = 3 # Radius of numerical mask to hide coro res = GreeDS(cube, angles, r=r, l=l, r_start=r_start, pup_size=pup_size)