Accueil > Actualités > Séminaires > Séminaire de Guillaume Charpiat


Titre : Deep Learning for Satellite Imagery: Semantic Segmentation, Non-Rigid Alignment, and Self-Denoising
Nom du conférencier : Guillaume Charpiat
Son affiliation : Équipe TAU, INRIA Saclay / LRI – Université Paris-Sud
Laboratoire organisateur : IPSL
Date et heure : 04-12-2019 10h30
Lieu : Campus Pierre and Marie Curie (Sorbonne Université) - 4 place Jussieu - Paris 5 - salle 105 du LIP6 couloir 25-26 1er étage
Résumé :

Le groupe IA et Climat ( ) qui réunit des chercheurs de l’IPSL et le LIP6 organise un séminaire public (sans inscription) bi-trimestriel pour présenter des recherches pluridisciplinaires concernant l’utilisation des méthodes d’intelligence Artificielle dans les sciences de l’environnement et du climat. Cette initiative est soutenue par le groupe de travail SAMA (Statistiques pour l’analyse, la modélisation et l’assimilation) de l’IPSL.

Neural networks have been producing impressive results in computer vision these last years, in image classification or segmentation in particular. To be transferred to remote sensing, this tool needs adaptation to its specifics: large images, many small objects per image, keeping high-resolution output, unreliable ground truth (usually mis-registered). We will review the work done in our group for remote sensing semantic segmentation, explaining the evolution of our neural net architecture design to face these challenges, and finally training a network to register binary cadaster maps to RGB images while detecting new buildings if any, in a multi-scale approach. We will show in particular that it is possible to train on noisy datasets, and to make predictions at an accuracy much better than the variance of the original noise. To explain this phenomenon, we build theoretical tools to express input similarity from the neural network point of view, and use them to quantify data redundancy and associated expected denoising effects.
If time permits, we might also present work on hurricane track forecast from reanalysis data (2-3D coverage of the Earth’s surface with temperature/pressure/etc. fields) using deep learning.

Notice Biographie

After a PhD thesis at ENS on shape statistics for image segmentation, and a year in Bernhard Schölkopf’s team at MPI Tübingen on kernel methods for medical imaging, Guillaume Charpiat joined INRIA Sophia-Antipolis to work on computer vision, and later INRIA Saclay to work on machine learning. Lately, he has been focusing on deep learning, with in particular remote sensing imagery as an application field.

Contact :