Descripteur
Termes IGN > informatique > intelligence artificielle > apprentissage automatique > apprentissage profond
apprentissage profond |
Documents disponibles dans cette catégorie (647)
![](./images/expand_all.gif)
![](./images/collapse_all.gif)
Etendre la recherche sur niveau(x) vers le bas
Titre : Learning stereo reconstruction with deep neural networks Type de document : Thèse/HDR Auteurs : Stepan Tulyakov, Auteur ; François Fleuret, Directeur de thèse ; Anton Ivanov, Directeur de thèse Editeur : Lausanne : Ecole Polytechnique Fédérale de Lausanne EPFL Année de publication : 2020 Importance : 139 p. Format : 21 x 30 cm Note générale : bibliographie
Thèse présentée à l'Ecole Polytechnique Fédérale de Lausanne pour l’obtention du grade de Docteur ès SciencesLangues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] apprentissage profond
[Termes IGN] classification semi-dirigée
[Termes IGN] contrainte géométrique
[Termes IGN] couple stéréoscopique
[Termes IGN] entropie
[Termes IGN] estimateur
[Termes IGN] étalonnage géométrique
[Termes IGN] modèle stéréoscopique
[Termes IGN] profondeur
[Termes IGN] réalité de terrain
[Termes IGN] reconstruction 3D
[Termes IGN] reconstruction d'image
[Termes IGN] vision par ordinateur
[Termes IGN] vision stéréoscopiqueRésumé : (auteur) Stereo reconstruction is a problem of recovering a 3d structure of a scene from a pair of images of the scene, acquired from different viewpoints. It has been investigated for decades and many successful methods were developed. The main drawback of these methods, is that they typically utilize a single depth cue, such as parallax, defocus blur or shading, and thus are not as robust as a human visual system that simultaneously relies on a range of monocular and binocular cues. This is mainly because it is hard to manually design a model, accounting for multiple depth cues. In this work, we address this problem by focusing on deep learning-based stereo methods that can discover a model for multiple depth cues directly from training data with ground truth depth. The complexity of deep learning-based methods, however, requires very large training sets with ground truth depth, which is often hard or costly to collect. Furthermore, even when training data is available it is often contaminated with noise, which reduces the effectiveness of supervised learning. In this work, in Chapter 3 we show that it is possible to alleviate this problem by using weakly supervised learning, that utilizes geometric constraints of the problem instead of ground truth depth. Besides the large training set requirement, deep stereo methods are not as application-friendlyas traditional methods. They have a large memory footprint and their disparity range is fixed at training time. For some applications, such as satellite stereo i magery, these are serious problems since satellite images are very large, often reaching tens of megapixels, and have a variable baseline, depending on a time difference between stereo images acquisition. In this work, in Chapter 4 we address these problems by introducing a novel network architecture with a bottleneck, capable of processing large images and utilizing more context, and an estimator that makes the network less sensitive to stereo matching ambiguities and applicable to any disparity range without re-training. Because deep learning-based methods discover depth cues directly from training data, they can be adapted to new data modalities without large modifications. In this work, in Chapter 5 we show that our method, developed for a conventional frame-based camera, can be used with a novel event-based camera, that has a higher dynamic range, smaller latency, and low power consumption. Instead of sampling intensity of all pixels with a fixed frequency, this camera asynchronously reports events of significant pixel intensity changes. To adopt our method to this new data modality, we propose a novel event sequence embedding module, that firstly aggregates information locally, across time, using a novel fully-connected layer for an irregularly sampled continuous domain, and then across discrete spatial domain. One interesting application of stereo is a reconstruction of a planet’s surface topography from satellite stereo images. In this work, in Chapter 6 we describe a geometric calibration method, as well as mosaicing and stereo reconstruction tools that we developed in the framework of the doctoral project for Color and Stereo Surface Imaging System onboard of ESA’s Trace Gas Orbiter, orbiting Mars. For the calibration, we propose a novel method, relying on starfield images because large focal lengths and complex optical distortion of the instrument forbid using standard methods. Scientific and practical results of this work are widely used by a scientific community. Note de contenu : 1- Introduction
2- Background
3- Weakly supervised learning of deep patch-matching cost
4- Applications-friendly deep stereo
5- Dense deep event-based stereo
6- Calibration of a satellite stereo system
7- ConclusionsNuméro de notice : 25795 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Note de thèse : Thèse de Doctorat : Sciences : Lausanne : 2020 En ligne : https://infoscience.epfl.ch/record/275342?ln=fr Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=95025
Titre : Low level feature detection in SAR images Type de document : Thèse/HDR Auteurs : Chenguang Liu, Auteur ; Florence Tupin, Directeur de thèse ; Yann Gousseau, Directeur de thèse Editeur : Paris [France] : Télécom ParisTech Année de publication : 2020 Importance : 138 p. Format : 21 x 30 cm Note générale : bibliographie
Thèse de doctorat de l’Institut Polytechnique de Paris préparée à Télécom Paris, Spécialité de doctorat : Signal, Images, Automatique et robotiqueLangues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image radar et applications
[Termes IGN] apprentissage profond
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] détection de contours
[Termes IGN] gradient
[Termes IGN] image radar moirée
[Termes IGN] modèle de Markov
[Termes IGN] segment de droiteIndex. décimale : THESE Thèses et HDR Résumé : (auteur) In this thesis we develop low level feature detectors for Synthetic Aperture Radar (SAR) images to facilitate the joint use of SAR and optical data. Line segments and edges are very important low level features in images which can be used for many applications like image analysis, image registration and object detection. Contrarily to the availability of many efficient low level feature detectors dedicated to optical images, there are very few efficient line segment detector and edge detector for SAR images mostly because of the strong multiplicative noise. In this thesis we develop a generic line segment detector and an efficient edge detector for SAR images.The proposed line segment detector which is named as LSDSAR, is based on a Markovian a contrario model and the Helmholtz principle, where line segments are validated according to their meaningfulness. More specifically, a line segment is validated if its expected number of occurences in a random image under the hypothesis of the Markovian a contrario model is small. Contrarily to the usual a contrario approaches, the Markovian a contrario model allows strong filtering in the gradient computation step, since dependencies between local orientations of neighbouring pixels are permitted thanks to the use of a first order Markov chain. The proposed Markovian a contrario model based line segment detector LSDSAR benefit from the accuracy and efficiency of the new definition of the background model, indeed, many true line segments in SAR images are detected with a control of the number of false detections. Moreover, very little parameter tuning is required in the practical applications of LSDSAR. The second work of this thesis is that we propose a deep learning based edge detector for SAR images. The contributions of the proposed edge detector are two fold: 1) under the hypothesis that both optical images and real SAR images can be divided into piecewise constant areas, we propose to simulate a SAR dataset using optical dataset; 2) we propose to train a classical CNN (convolutional neural network) edge detector, HED, directly on the graident fields of images. This, by using an adequate method to compute the gradient, enables SAR images at test time to have statistics similar to the training set as inputs to the network. More precisely, the gradient distribution for all homogeneous areas are the same and the gradient distribution for two homogeneous areas across boundaries depends only on the ratio of their mean intensity values. The proposed method, GRHED, significantly improves the state-of-the-art, especially in very noisy cases such as 1-look images. Note de contenu : 1- Context
2- SAR basics, statistics of SAR images and data used in this thesis
I Line segment detection in SAR images
3- Introduction
4- LSD, a line segment detector with false detection control
5- LSDSAR, a generic line segment detector for SAR images
6- Experiments
II Edge detection in SAR images using CNNs
7- Introduction
8- Presentation of the HED method and of the training dataset
9- GRHED, introducing a hand-crafted layer before the usual CNNs
10- Experiments
11- Summary of the thesisNuméro de notice : 25878 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse française Note de thèse : Thèse de Doctorat : Spécialité : Signal, Images, Automatique et robotique : Paris : 2020 nature-HAL : Thèse DOI : sans En ligne : https://tel.archives-ouvertes.fr/tel-02861903/document Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=95689
Titre : Multi-agent systems : Strategies and applications Type de document : Monographie Auteurs : Ricardo Lopez-Ruiz, Éditeur scientifique Editeur : London [UK] : IntechOpen Année de publication : 2020 Importance : 170 p. Format : 19 x 27 cm ISBN/ISSN/EAN : 978-1-78985-394-0 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Intelligence artificielle
[Termes IGN] apprentissage automatique
[Termes IGN] apprentissage par renforcement
[Termes IGN] reconstruction 3D
[Termes IGN] reconstruction d'image
[Termes IGN] réseau neuronal artificiel
[Termes IGN] système multi-agentsRésumé : (éditeur) Research on multi-agent systems is enlarging our future technical capabilities as humans and as an intelligent society. During recent years many effective applications have been implemented and are part of our daily life. These applications have agent-based models and methods as an important ingredient. Markets, finance world, robotics, medical technology, social negotiation, video games, big-data science, etc. are some of the branches where the knowledge gained through multi-agent simulations is necessary and where new software engineering tools are continuously created and tested in order to reach an effective technology transfer to impact our lives. This book brings together researchers working in several fields that cover the techniques, the challenges and the applications of multi-agent systems in a wide variety of aspects related to learning algorithms for different devices such as vehicles, robots and drones, computational optimization to reach a more efficient energy distribution in power grids and the use of social networks and decision strategies applied to the smart learning and education environments in emergent countries. We hope that this book can be useful and become a guide or reference to an audience interested in the developments and applications of multi-agent systems. Note de contenu : 1- Cooperative adaptive learning control for a group of nonholonomic UGVs by output feedback
2- Multiagent systems for 3D reconstruction applications
3- A Q-learning-based approach for simple and multi-agent systems
4- Multi-Agent systems, simulation and nanotechnology
5- Applications of multi-agent system in power system engineering
6- Architecture of a microgrid and optimal energy management system
7- Multi-agent systems based advanced energy management of smart micro-grid
8- Smart learning environment: Paradigm shift for onlint learning
9- ICT: Vehicle for educational development and social TransformationNuméro de notice : 28572 Affiliation des auteurs : non IGN Thématique : INFORMATIQUE Nature : Recueil / ouvrage collectif DOI : 10.5772/intechopen.81766 Date de publication en ligne : 22/04/2020 En ligne : https://doi.org/10.5772/intechopen.81766 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97871
Titre : Multi-sensor information fusion Type de document : Monographie Auteurs : Xue-Bo Jin, Éditeur scientifique ; Yuan Gao, Éditeur scientifique Editeur : Bâle [Suisse] : Multidisciplinary Digital Publishing Institute MDPI Année de publication : 2020 Importance : 602 p. ISBN/ISSN/EAN : 978-3-03928-303-3 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Acquisition d'image(s) et de donnée(s)
[Termes IGN] apprentissage profond
[Termes IGN] classification semi-dirigée
[Termes IGN] filtre de Kalman
[Termes IGN] fusion de données multisource
[Termes IGN] série temporelle
[Termes IGN] temps réelRésumé : (Editeur) This book includes papers from the section “Multisensor Information Fusion”, from Sensors between 2018 to 2019. It focuses on the latest research results of current multi-sensor fusion technologies and represents the latest research trends, including traditional information fusion technologies, estimation and filtering, and the latest research, artificial intelligence involving deep learning. Numéro de notice : 26505 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Recueil / ouvrage collectif DOI : 10.3390/books978-3-03928-303-3 En ligne : https://doi.org/10.3390/books978-3-03928-303-3 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97083 Past and future evolution of French Alpine glaciers in a changing climate: a deep learning glacio-hydrological modelling approach / Jordi Bolibar Navarro (2020)
![]()
Titre : Past and future evolution of French Alpine glaciers in a changing climate: a deep learning glacio-hydrological modelling approach Type de document : Thèse/HDR Auteurs : Jordi Bolibar Navarro, Auteur ; Antoine Rabatel, Auteur ; Isabelle Gouttevin, Auteur ; Eric Sauquet, Auteur Editeur : Grenoble [France] : Université Grenoble Alpes Année de publication : 2020 Importance : 143 p. Format : 21 x 30 cm Note générale : bibliographie
Thèse pour obtenir le grade de Docteur de l'Université Grenoble Alpes, Spécialité : Sciences de la Terre et de l’Univers et de l’EnvironnementLangues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géomatique
[Termes IGN] Alpes (France)
[Termes IGN] analyse diachronique
[Termes IGN] apprentissage automatique
[Termes IGN] bilan de masse
[Termes IGN] changement climatique
[Termes IGN] glacier
[Termes IGN] modèle de simulation
[Termes IGN] modèle hydrographique
[Termes IGN] modèle numérique de surface
[Termes IGN] réseau neuronal artificielIndex. décimale : THESE Thèses et HDR Résumé : (auteur) The European Alps are among the most affected regions in the world by climate change, displaying some of the strongest glacier retreat rates. Long-term interactions between society, mountain ecosystems and glaciers in the region raise important questions on the future evolution of glaciers and their derived environmental and socioeconomical impacts. In order to correctly assess the regional response of glaciers in the French Alps to climate change, there is a need for adequate modelling tools. In this work, we explore new ways to tackle both glacier evolution and glacio-hydrological modelling at a regional scale. Glacier evolution modelling has traditionally been performed using empirical or physical approaches, which are becoming increasingly challenging to optimize with the ever growing amount of available data. Here, we present, to our knowledge, the first effort ever to apply deep learning (i.e. deep artificial neural networks) to simulate the evolution of glaciers. Since both the climate and glacier systems are highly nonlinear, traditional linear mass balance models offer a limited representation of climate-glacier interactions. We show how important nonlinearities in glacier mass balance are captured by deep learning, substantially improving model performance over linear methods.This novel method was first applied in a study to reconstruct annual mass balance changes for all glaciers in the French Alps for the 1967-2015 period. Using climate reanalyses, topographical data and glacier inventories, we demonstrate how such an approach can be successfully used to reconstruct large-scale mass balance changes from observations. This study also offered new insights on how glaciers evolved in the French Alps during the last half century, confirming the rather neutral observed mass balance rates in the 1980s and displaying a well-marked acceleration in mass loss from the 2000s onwards. Important differences between regions are found, with the Mont-Blanc massif presenting the lowest mass loss and the Chablais being the most affected one. Secondly, we applied this modelling framework to simulate the future evolution of all glaciers in the region under multiple (N=29) climate change scenarios. Our estimates indicate that most ice volume in the region will be lost by the end of the 21st century independently from future climate scenarios. We predict average glacier volume losses of 74%, 80% and 88% under RCP 2.6 (n=3), RCP 4.5 (n=13) and RCP 8.5 (n=13), respectively. By the end of the 21st century the French Alps will be largely ice-free, with glaciers only remaining in the Mont-Blanc and Pelvoux massifs. Note de contenu : Introduction
1- Glaciers
2- Glacierized mountain catchments
3- OutlookNuméro de notice : 28311 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Thèse française Note de thèse : Thèse de Doctorat : Sciences de la Terre et de l’Univers et de l’Environnement : Grenoble : 2020 Organisme de stage : Institut des Géosciences de l’Environnement (Grenoble) DOI : sans En ligne : https://tel.archives-ouvertes.fr/tel-03052063v2/document Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98202 PermalinkPermalinkPermalinkPermalinkPermalinkPermalinkSuperpixel-enhanced deep neural forest for remote sensing image semantic segmentation / Li Mi in ISPRS Journal of photogrammetry and remote sensing, vol 159 (January 2020)
PermalinkSystème de traitement d’images temps réel dédié à la mesure de champs denses de déplacements et de déformations / Seyfeddine Boukhtache (2020)
PermalinkTorch-Points3D: A modular multi-task framework for reproducible deep learning on 3D point clouds / Thomas Chaton (2020)
PermalinkUnderwater field equipment of a network of landmarks optimized for automatic detection by AI / Laurent Beaudoin (2020)
PermalinkUnsupervised satellite image time series analysis using deep learning techniques / Ekaterina Kalinicheva (2020)
PermalinkVers une occupation du sol France entière par imagerie satellite à très haute résolution / Tristan Postadjian (2020)
PermalinkPermalinkShip identification and characterization in Sentinel-1 SAR images with multi-task deep learning / Clément Dechesne in Remote sensing, Vol 11 n° 24 (December-2 2019)
PermalinkCombining Sentinel-1 and Sentinel-2 Satellite image time series for land cover mapping via a multi-source deep learning architecture / Dino Lenco in ISPRS Journal of photogrammetry and remote sensing, Vol 158 (December 2019)
PermalinkDeep learning for conifer/deciduous classification of airborne LiDAR 3D point clouds representing individual trees / Hamid Hamraz in ISPRS Journal of photogrammetry and remote sensing, Vol 158 (December 2019)
PermalinkHalf a percent of labels is enough: efficient animal detection in UAV imagery using deep CNNs and active learning / Benjamin Kellenberger in IEEE Transactions on geoscience and remote sensing, vol 57 n° 12 (December 2019)
PermalinkMatching of TerraSAR-X derived ground control points to optical image patches using deep learning / Tatjana Bürgmann in ISPRS Journal of photogrammetry and remote sensing, Vol 158 (December 2019)
PermalinkComparison between convolutional neural networks and random forest for local climate zone classification in mega urban areas using Landsat images / Cheolhee Yoo in ISPRS Journal of photogrammetry and remote sensing, vol 157 (November 2019)
PermalinkContext pyramidal network for stereo matching regularized by disparity gradients / Junhua Kang in ISPRS Journal of photogrammetry and remote sensing, vol 157 (November 2019)
Permalink