Related Publications by Category
Data Descriptor
Contini, M., Illien, V., Julien, M. et al. Seatizen Atlas: a collaborative dataset of underwater and aerial marine imagery. Sci Data 12, 67 (2025).
Gogendeau, P. et al. An autonomous surface vehicle for acoustic tracking, bathymetric and photogrammetric surveys, 2025.
Artificial Intelligence
Contini, M., Illien, V., Poulain, S., Bernard, S., Barde, J., Bonhommeau, S., & Joly, A. (2025). The point is the mask: scaling coral reef segmentation with weak supervision. arXiv preprint arXiv:2508.18958.
Contini, Matteo, et al. From underwater to drone: A novel multi-scale knowledge distillation approach for coral reef monitoring. Ecological Informatics (2025): 103149.
Talpaert Daudon, Justine, et al. Geoai for marine ecosystem monitoring: A complete workflow to generate maps from ai model predictions. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. 48. Copernicus GmbH, 2023.
DOI: https://doi.org/10.5194/isprs-archives-XLVIII-4-W7-2023-223-2023
eDNA
Cowart DA, Chevrier T, Nieblas A-E, Chanut J, Bernard S and Bonhommeau S (2024) Detecting local variations across metazoan communities in back-reef depressions of Reunion Island (Mascarene Archipelago) through environmental DNA survey. Front. Mar. Sci. 11:1423676
Join the Seatizen Atlas Community
Discover, visualize, and download the data produced by the Seatizen Team. Access georeferenced frames, habitat predictions, and more.
Join usSeatizen Workflow
The Seatizen workflow manages every step from data acquisition to publication, in full accordance with the FAIR principles. Raw data are organized, processed, and versioned on Zenodo with a DOI, ensuring transparency, traceability, and reproducible visualizations. All source code is hosted on GitHub, along with a Docker image to guarantee consistent execution and reusability across environments.
plancha-workflow
Convert raw data into georeferenced frames with centimeter-level accuracy using PPK, and enrich each frame with relevant metadata.
View Repositorydrone-workflow
Extract metadata from each frame and produce deliverables including GeoPackages, PDFs, and additional formats.
View Repositoryplancha-inference
Use Jacques to filter irrelevant images, predict multiple coral classes per image, and create PDF previews and raster files for analysis.
View Repositorydrone-upscaling
Build a training dataset and train a model to perform multilabel classification on drone orthophotos based on ASV prediction rasters.
View Repositorythe-point-is-the-mask
Configure, train, and run inference on drone orthophotos using predictions from an ASV to achieve weakly supervised semantic segmentation.
View Repositoryfiftyone-tools
Tools to manually annotated images with labels to enhance new dataset using fiftyone software
View Repositoryzenodo-tools
Tools to communicate with zenodo-api:• zenodo-upload: Send a session on zenodo (raw data, processed data, custom data, update metadata, delete draft session)• zenodo-download: Download a session from a doi or a session name matching the good doi• zenodo-manager: Create or update a geopackage to centralize all metadata and generate subtiles to easily manipulate all sessions
View Repositoryseatizen-monitoring
Explore data produced by the Seatizen Team and stored in the Seatizen Atlas database. Select the frames you’re interested in to generate custom training datasets.
View Repositorycog-server
Visualize orthophoto on qgis.Serve ortho, bathy, habitat map prediction as raster under one URL.Can get depth at one point or habitat map prediction.
View Repository