About me
I’m currently working as a PhD Candidate in the SIROCCO team advised by Christine GUILLEMOT and Aline ROUMY. My PhD focuses on low-light image restoration in the context of autonomous vehicles for instance. I’m also a self-taught guitar player in my free time and I practice rock climbing & bouldering.
Interests
- Computer Vision & Signal Processing
- Deep Learning
- Generative modeling
- Optimal transport
- Computational biology
- Mathematical physics
- Dynamical systems
- FOSS
Education
-
PhD Candidate in Signal Processing & Computer Vision, 2023
Université Rennes 1
-
MSc in Machine Learning (French degree in engineering), 2020
ESIEE Paris
-
Exchange semester in Machine Learning, 2018
Polytechnique Montréal
Recent Publications
(2023).
Illumination colorée et information mutuelle pour le modèle Retinex en restauration d'images à faible luminosité.
In GRETSI 2023 (Oral).
(2023).
GAN architecture leveraging a retinex model with colored illumination for low-light image restoration.
In IEEE Access.
(2022).
A New Regularization for Retinex Decomposition of Low-Light Images.
In ICIP 2022 (Oral).
(2020).
Sexing Hyalella Azteca with machine learning for bioaccumulation test.
In SETAC SciCon 2.
Work Experience
PhD Candidate
- Member of the SIROCCO team (Analysis representation, compression and communication of visual data).
- Thesis under supervision of Christine GUILLEMOT and Aline ROUMY.
- Developed new deep learning-based methods to restore low-light images based on the Retinex theory (correlated source separation + underdetermined problem).
- Explored a wild range of topics (Generative modeling, GANs, diffusion and flow-based models, optimal transport and Schrödinger Bridge, Contrastive learning, Camera image processing pipeline, Classical image processing algorithms, …).
- Used different real‐world datasets such as the Waymo dataset and satellite images provided by Airbus Defence and Space.
- Part of a joint project with Airbus Defence and Space.
- Reviewer for IEEE Transactions on Image Processing.
Research Intern
- Implemented a Deep Learning model capable of sexing hyalellas reaching 99.2% accuracy.
- Created the dataset with 2000+ pictures.
- Internship done as part of the sustainable development lab working on ecotoxicology.
- Applied a preprocessing on the images (mathematical morphology, pattern recognition, feature extraction, …).
- Co‐supervision of a group of 5 undergraduate students.
- In partnership with the Gaspard‐Monge laboratory.
Research Intern
- Developed a platform to collect and share electricity, water and gas usage data for the creation of the new Gustave Eiffel University.
- Applied data clustering algorithms to characterize the consumption of the different buildings.
- Identified 4 types of electricity consumption behaviors with the intention of saving energy.
- Conducted tests with different sensors to check the feasibility of a local low‐cost solution (Internet of things & LoRaWAN).
- Communicated with several providers to get quotations with the aim of setting a crowd flow forecasting system (counting pedestrians and vehicles).