Programme
14h - 14h40 Mihai Datcu
The Earth Observation Sensory Gap: from Bayesian Inference to Deep Learning
Abstract: The sensory gap is the dissimilarity between the actual nature of an object and the information extracted from the signals recorded by a sensor observing this object. Earth Observation (EO) images are sensor records, gathering the signature of the observed scenes in a specific electromagnetic spectrum, therefore an indirect signature of the imaged scenes. The challenge is in inverting the physically meaningful parameters of the scene from these observations. The lecture is presenting a communication channel model for the parameter retrieval problem. The source is the ensemble of EO data, and the information contained in the data is a message. Modeling the data processing chain as a communication channel allows measuring and quantifying the amount of information each feature descriptor can provide about a set of images. The generative Bayesian models, as Latent Dirichlet Allocation (LDA), Restricted Boltzmann Machine, Generative Adversarial Networks and the latest Deep Learning paradigms are discussed and exemplified as solutions for the EO sensory gap.
14h45 - 15h30 Yuliya Tarabalka
Can we classify the world? Where Deep Learning Meets Remote Sensing
Abstract: Deep learning has been recently gaining significant attention for the analysis of data in multiple domains. It seeks to model high-level knowledge as a hierarchy of concepts. With the exploding amount of available data, the improvement of hardware and the advances in training methodologies, now such hierarchies can contain many more processing layers than before, hence the adoption of the term "deep".
In remote sensing, recent years have witnessed a remarkable increase in the amount of available data, due to a consistent improvement in the spectral, spatial and temporal resolutions of the sensors. Moreover, there are new sources of large-scale open access imagery, governments are releasing their geographic data to the public, and collaborative platforms are producing impressive amounts of cartography. With such an overwhelming amount of information, it is of paramount importance to develop smart systems that are able to handle and analyze these data. The scalability of deep learning and its ability to gain insight from large-scale datasets, makes it particularly interesting to the remote sensing community. It is often the case, however, that the deep learning advances in other domains cannot be directly applied to remote sensing. The type of input data and the constraints of remote sensing problems require the design of specific deep learning techniques.
In this talk, I will discuss how deep learning approaches help in remote sensing image interpretation. In particular, I will focus on the most powerful architectures for semantic labeling of aerial and satellite optical images, with the final purpose to produce and update world maps.
15h30 - 16h Pause
16h - 16h40 Session Poster
Correcting misaligned buildings over aerial images by a Deep Learning multi-resolution approach
Nicolas Girard, Yuliya Tarabalka, Guillaume Charpiat
Generative Adversarial Network (GAN) for Remote Sensing Images unsupervised Learning
Amina Ben Hamida, Alexandre Benoît, Patrick Lambert, Chokri Ben Hamar
Investigating patch-based and pixel-based deep architectures for dense semantic segmentation
Maria Papadomanolaki, Maria Vakalopoulou, Konstantinos Karantzalos
Fully Convolutional Siamese Networks for Change Detection
Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch, Yann Gousseau
Joint Learning from Earth Observation and OpenStreetMap Data to Get Faster Better Semantic Maps
Nicolas Audebert, Bertrand Le Saux, Sébastien Lefèvre
CLOUD-GAN: Cloud Removal for Sentinel-2 Imagery using Generative Adversarial Networks
Praveer Singh, Nikos Komodakis
Démonstration du potentiel des séries temporelles SAR pour la détection d'activité à grande échelle
Élise Colin-Koeniguer
Apprentissage actif pour l’annotation d’images aériennes appliqué aux suivis environnementaux
Mathieu Laroze, Romain Dambreville, Chloé Friguet, Ewa Kijak, Sébastien Lefèvre