Research
Current focuses:
Label-efficient learning
Representation learning
Self-supervised/Weakly-supervised learning
Few-shot learning
Multi-task learning
Knowledge distillation
Federated learning
Vision and EO tasks: classification, object detection, super-resolution
Research interests:
Computer vision, image analysis and processing
Machine learning, deep learning
Signal and image processing on graphs
Morphological mathematics on images
Hierarchical learning and structured classification
Remote sensing data fusion (optical, radar, lidar, sonar)
Remote sensing applications (detection, segmentation, super-resolution)
Research projects:
-----On-going---
AXOLOTL (2024-2027) - Role: Partner's Lead
From sky to seafloor observation: Achieving excellence in oceanic surveillance and conservation through deep learning.
Funding: 1.5 M€, HORIZON EUROPE
Consortium: CMMI Cyprus (main PI), UBS France and VLIZ Belgium
ROMMEO (2022-2024) - Principal Investigator
Robust multitask learning via mutual knowledge distillation for earth observation
Funding: 63k€, Région Bretagne (France)
OTTOPIA (2021-2024)
Earth observation by optimal transport for artificial intelligence
Funding: 467k€, ANR
-----Past Projects----
OWFSOMM (2020-2023)
Offshore wind farm surveys of marine megafauna: standardization of tools and methods for monitoring at OWF scales
Funding: 529K€, ANR/FEM
SEMMACAPE (2019-2023)
Monitoring and study of marine megafauna by automatic characterization in wind farms
Funding: 128k€, ADEME
ANR MULTISCALE (2019-2023)
Multi-variate, -temporal, -resolution and -source remote sensing image analysis and learning
ANR DEEPDETECT (2018-2020)
Detection and recognition of multiple objects on variable backgrounds by deep learning
R&T CNES CARS (2018-2019)
Vehicle counting by deep learning from Pléiades images
DELORA (2016-2018)
Augmented reality detection and relocation of buried networks
Editorial/Expertise:
Expertises
Member of Editorial Board of Digital Signal Processing, Elsevier (since 2022)
Evaluation member of CE23 – Intelligence Artificielle », ANR Appel à projets générique 2021.
Mentor and Evaluator of IT Swin Hackaton 2023 (Proptech/EduTech/HealthTech), Swinburne University.
Applied Research Specialist at VinAI Research (2020-2021)
Member of PhD Committee (Jury de thèse)
Sara Akodad, Université de Bordeaux, IMS Lab. PhD topic : « Ensemble learning methods on the space of covariance matrices : application to remote sensing scene and multivariate time series classification ». defended on 08/11/2021
Member of Individual PhD Mornitoring Committee (Comité de suivi individuel, CSI)
Kévin Planolles, Université Montpellier, LIRMM Lab. PhD topic : « automated monitoring of fish biodiversity with work on detection algorithms », started 09/2022.
Shuo Yang, CY Cergy Paris Université. PhD topic : « Graph neural networks for human pose estimation », started 05/2023.
Guest Editor
Special issue on Machine learning for earth observation data, Machine Learning, with Thomas Corpetti (LETG-Rennes UMR 6554, Rennes), Dino Ienco (INRAE, UMR TETIS, Montpellier), Sébastien Lefèvre (IRISA, UMR 6074, Vannes) et Roberto Interdonato (INRAE, UMR TETIS, Montpellier).
Special issue on Feature-based methods for remote sensing image classification, Remote Sensing, with Carlos Lopez-Martinez (Universitat Politècnica de Catalunya, Espagne) et Pelich Ramona (Luxembourg Institute of Science and Technology, Luxembourg).
Co-chair of the MACLEAN (Machine learning for Earth Observation) workshop at ECML/PKDD, annually since 2019.
Co-chair of the MACLEAN session at GDR MADICS (2019-2022)