|
|
|
Programme > Workshop MACLEAN
MACLEAN @ Cap/RFIAP Workshop: Machine learning and computer vision in earth observation: scientific results versus industrial needs
With the increasing number of satellite images and data available and the increasing acquisition frequencies, the automatic interpretation of remote sensing and Earth observation data is a very active field. Sensors are now capable of providing (very) high resolution images with acquisition frequencies never before achieved. Due to this technological revolution, it is increasingly difficult to design methods capable of efficiently processing these data and taking advantage of the complementarities of the different data sources or the opportunities offered by the time series of satellite images. In this context, modern machine learning and computer vision methods can play a fundamental role.
In the framework of the CAp and RFIAP conferences, a workshop on machine learning and computer vision issues in the context of earth observation is planned for Tuesday 5 July. This workshop will be organised with the support of the MACLEAN action of the GDR MADICS as well as the SIG BreTel, which aims to bring together the environmental and data science communities. More precisely, the objective of the day will be to cross-reference the needs and expectations of industrialists in the field with the work of academic research laboratories. In doing so, the workshop will aim to raise awareness of the potential of the latest academic scientific developments, to confront them with industrial realities, but also to identify scientific issues that companies are facing and for which research work needs to be undertaken.
The day will consist of invited presentations (academic and industrial), doctoral/postdoctoral pitches, and a panel session.
THE TENTATIVE PROGRAMME IS AS FOLLOWS :
| 9:00 | Welcome | |
| 9:30 | Opening | Sébastien Lefèvre (IRISA / UBS, chair), Nicolas Bellec (IRISPACE coordinator), Jean-Marc Delvit (CNES), Hélène Bonfils (PCN Espace) |
| 11:00 | Keynote 1 | Pierre Allain (WIPSEA) and Minh-Tan Pham (IRISA / UBS) - From supervised to unsupervised deep learning for automatic detection of marine megafauna |
| 11:45 | Keynote 2 | Aurélien Colin (CLS & LabSTICC / IMTA) - Semantic segmentation of meteorological and oceanic phenomena on SAR images |
| 12:30 | Cocktail | |
| 14:00 | Keynote 3 | Iris de Gélis (Magellium & IRISA) - Deep learning based change detection in 3D point clouds |
| 14:45 | Doctoral/post-doctoral pitchs | Zoé Bessin (LETG) - Contribution of multi-angular satellite imagery to the monitoring and understanding of coastal cliff erosion Manal Hamzaoui (IRISA) - A Hierarchical Prototypical Network for Few-Shot Remote Sensing Scene Classification Gabriel Kasmi (RTE France / Mines Paris - PSL) - Leveraging earth observation data and deep learning to estimate the PV output in France Arthur Filoche (LIP6) - Variational assimilation of ocean observation with deep prior Diego Di Carlo (LETG) - Physics-informed neural network for super-resolution of turbulent flows Jimmy Daynac (LPG) - Contribution of Artificial Intelligence to cartography and analysis the dunes |
| 16:30 | Panel | Sébastien Lefèvre (moderator), Nicolas Bellec (IRISPACE), Jean-Marc Delvit (CNES), Frank de Morsier (Picterra, Switzerland), Nicolas Courty (IRISA / UBS) |
| 17:00 | Closing |
REGISTRATION
They are accessible through this link : https://framaforms.org/workshop-maclean-mardi-5-juillet-1654693360
| Personnes connectées : 3 | Vie privée |
|
