About me
I am an Innovation Project Manager at French Mapping Agency IGN and researcher at LASTIG working on change detection for earth observation and deep learning applied to geospatial data.
Previously I was a PostDoctoral researcher in the IMAGINE team of the ENPC, working with Mathieu Aubry on unsupervised learning for text recognition.
I did my Ph.D. at Université Paris Saclay and Télécom Paris, in the IMAGE team. I worked under the supervision of Yann Gousseau, Saïd Ladjal and Olivier Bonfait on deep learning for art analysis. My PhD manuscript can be found here.
My research interests include deep learning applied for geospatial data and historical images.
👨🔬 Research areas
- Change Detection
- Representation learning
- Multimodal Learning
- Earth Observation
🥼 Research projects
👹 OGRE - Observation de la Terre et IA Générative pour la Reconnaissance d’Événements rares (2026-2029)
Website to be published
Partners: LASTIG (Univ. Gustave Eiffel, IGN-ENSG), CEDRIC (Cnam Paris), DTIS (ONERA), ISIR (Sorbonne Université, CNRS).
Participants: Nicolas Audebert (PI), Nicolas Gonthier, Clément Mallet, Arnaud Breloy, Javiera Castillo, Flora Weissgerber, Adrien Chan Hon Tong, Clément Rambour, Alasdair Newson.
Funding: Agence Innovation Défense
Earth Observation and Generative models for Rare Events detection (OGRE) explores generative models and their application to extreme events detection in multimodal remote sensing data (optical and SAR). It investigates how to leverage large generative models as likelihood estimators, to detect finely-localized anomalies, both in time and space, and to perform change detection in satellite image time series. It targets a broad range of applications from flood surveillance, ice melting monitoring and urban growth analysis.
✈️ Change Detection in Very High Resolution Images (2022-2025)
IGN PhD Funding
🌍 Multimodal Self-Supervised Learning for Earth Observation (2022-2026)
CNES and IGN PhD Funding