Projects

Here is a list of of the various projects I have been involved in and which funded my research.

As leader

Ongoing

2023 - 2026 FORMAL “From observing to modelling ocean life” (Sorbonne Université, Fr; 500k€, coPI). Bring together oceanographers and applied mathematicians to foster improved use of data and models of the marine environment. Involves six kinds of data (omics, imaging, biologging, acoustics, biogeochemical profiles, satellite images), four types of models (biogeochemical, niche, interaction, lagrangian), and leverages machine learning techniques for data processing, analysis, assimilation into models.

Past

2022 - 2023 LOVNOWER “Machine learning and computer vision for plankton imagery” (France Relance, Fr; 220k€, PI).
Partnership with a computer vision company to develop machine learning-based tools for classification and visual search in the EcoTaxa plankton imagery platform.

2019 - 2023 WWWPIC World Wide Web of Plankton Image Curation (Belmont Forum; 1M€, PI).
Setup an international infrastructure for the classification of plankton images assisted by machine learning and the sharing of the resulting datasets; use this infrastructure to tackle questions about the distribution of plankton from the mesoscale to the global ocean.

2019 - 2021 ARTIFACTZ “Application de l’intelligence artificielle à l’identification de traits fonctionnels à partir d’images à haute résolution” (Sentinelle-Nord; 40k€, coPI).
Use artificial intelligence to extract ecological traits from plankton images.

2019 EMODnet Biology Data Grant “Plankton Imaging Data” (EU; 20k€ PI).
Use the EcoTaxa imagery management platform to send large datasets of plankton concentrations at various locations and dates around the world; in total, over 4M occurrences were reported.

2017 - 2018 DL-PIC “Deep Learning for Plankton Image Classification” (CNRS; 23k€, PI).
Use dedicated Convolutional Neural Networks for plankton images classification.

2012 - 2016 Biophysical interactions at mesoscale in the ocean (PUF; 240k$, PI).
Designed and run this research, innovation, and education program, with the University of Miami. The program funded three research actions (dealing with plankton and fish larvae), the evolution of two instruments, and directly involved 12 professors and 26 students.

As contributor

Ongoing

2023 - 2026 Blue-Cloud 2026 “A federated European FAIR and Open Research Ecosystem for oceans, seas, coastal and inland waters”: (Horizon EU; 360k€, PI for Sorbonne Université).
Integrate plankton imagery data as a FAIR data source into international data infrastructure; leverage these infrastructures and cloud computing to define operational species distribution modelling pipeline and map biodiversity at global scale.

DTO-Bioflow, WP3 “Enabling sustained flows of biodiversity monitoring data into the DTO”: (Horizon EU; 445k€, PI for Sorbonne Université).
Define and implement automated flows of data from plankton imagery instruments to the Digital Twin of the Ocean. Exploit this data to better constrain carbon flux at global scale.

ANERIS “operAtional seNsing lifE technologies for maRIne ecosystemS” WP3 Imagery (Horizon EU; 170k€, participant).
Deploy imaging instruments on moored platforms and process their data on-the-fly to study high frequency changes in marine plankton ecosystems.

BiOcean5D, WP3 “Data to knowledge, a digital foundation for holistic marine biodiversity assessment” (Horizon EU; 250k€, task PI).
Study biodiversity along coast to offshore gradients along the coast of Europe through the Tara Europa cruise, in association with the coastal TREC sampling campaign of EMBL. Beyond scientific exploitation of the data, my role is also to ensure the proper flow of plankton imagery data towards international databases.

2022 - 2028 FUTURE-OBS “Observatoire augmenté pour les socio-ecosystèmes côtiers” (PPR Océans, Fr; 100k€, participant).
Integrate new ocean monitoring approaches (based on meta-omics and imagery) into the existing framework of French coastal monitoring systems. Our contribution is mostly around best practices, standards and training in plankton imagery.

2022 - 2026 Smart-Biodiv “Smart AI technologies for Biodiversity Research” (ANR, Fr; 88k€, PI for LOV).
Derive improved or new biodiversity indicators from existing biological data, using Artificial Intelligence techniques. Our contribution is more targeted towards providing (i) plankton imagery data from time series in the Mediterranean Sea and cruises in the Pacific and (ii) expertise regarding the ecological processes associated.

TraitZoo “Trait biogeography and functional diversity of marine mesozooplankton” (ANR, Fr; 134k€, PI for LOV).
Extract traits (size, pigment concentration, reproduction potential, energy reserves) from images of plankton and use this data to describe the biogeography and functional implications of those traits.

2022 - 2025 iMagine “Imaging data and services for aquatic science” (Horizon EU; 360k€, PI for Sorbonne Université).
Develop a new version of the image processing pipeline for the ZooScan instrument, using artificial intelligence; release this as a cloud service; leverage it to process a 50 years long time series of plankton imagery data and analyse it to detect potential effects of climate change.

PSSdb “A Pelagic Size Structure database” (NOAA, US; 145k€, participant).
Contribute to assembling a global database of concentrations of plankton and marine snow particles, organised per size class, from various imaging instruments, to calibrate carbon flux estimates in numerical models.

PopNCo, WP5 “Data integration and synthesis” (ANR, Fr; 130k€, WP PI).
Assess the spatio-temporal variability of connectivity between biological populations from genetic and modelling data, through graph-based approaches.

2020 - 2024 Theme 2 “Imaging and optics” of TechOceanS (H2020, EU; 300k€, participant).
Investigate image compression through auto-encoders and unsupervised image selection for optimal use of low-bandwidth transmission from autonomous in situ plankton imagers.

JERICO-S3, Tasks 6.3.1 “Biological imagery data” and 7.5 “JERICO infrastructure” (H2020, EU; 62k€, task PI).
Define international standards for quantitative imagery of plankton, implement those standards in data processing pipelines from several instruments, provide access to an infrastructure hosting the resulting data.

AtlantECO (H2020, EU; 300k€, participant).
Make a census of and map biodiversity and biology-driven carbon flux in the Atlantic Ocean through traditional and novel tools (genomics, imagery).

2015 - 2020

2019 - 2023 BlueCloud, Task 2.3 “Plankton demonstrator” (H2020, EU; 250k€, PI for Sorbonne Université).
Use genomics and images jointly to infer new species and functions for unassigned genomic reads from plankton samples; then predict the geographical distribution of these plankton groups, traits, and functions.

2019 - 2021 CARDINAL”Calanus redness index from artificial intelligence: applications to image analysis” (Sentinelle-Nord, The Arctic University of Norway; 154k€, participant).
Quantify redness of individual Calanus copepods from images and relate it to redness of surface waters seen from satellites.

2015 - 2019 BRIDGES, Task”Underwater Vision Profiler version 6” (H2020, EU; 450k€, participant). Develop an intelligent plankton camera for autonomous instruments; I lead the machine learning developments for onboard classification of images.

Before 2015

2013 - 2015 PERSEUS, Task 1.6”Ecoregionalisation” (H2020, EU; 100k€, task PI).
Use ecological niche modelling and multivariate clustering to define homogeneous regions within the Mediterranean Sea.

2011 - 2016 MERMEX, WP5 “Marine regionalisation planning” (CNRS; 60k€, WP PI).
Coordinated the data consolidation and valorisation activities of this structuring program; this involved biogeography studies as well as joint research with environmental economists.