Descripteur
Documents disponibles dans cette catégorie (1401)
Ajouter le résultat dans votre panier
Visionner les documents numériques
Affiner la recherche Interroger des sources externes
Etendre la recherche sur niveau(x) vers le bas
LANet: Local attention embedding to improve the semantic segmentation of remote sensing images / Lei Ding in IEEE Transactions on geoscience and remote sensing, vol 59 n° 1 (January 2021)
[article]
Titre : LANet: Local attention embedding to improve the semantic segmentation of remote sensing images Type de document : Article/Communication Auteurs : Lei Ding, Auteur ; Hao Tang, Auteur ; Lorenzo Bruzzone, Auteur Année de publication : 2021 Article en page(s) : pp 426 - 435 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] analyse de données
[Termes IGN] apprentissage profond
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] décodage
[Termes IGN] distribution spatiale
[Termes IGN] extraction de traits caractéristiques
[Termes IGN] segmentation sémantiqueRésumé : (auteur) The trade-off between feature representation power and spatial localization accuracy is crucial for the dense classification/semantic segmentation of remote sensing images (RSIs). High-level features extracted from the late layers of a neural network are rich in semantic information, yet have blurred spatial details; low-level features extracted from the early layers of a network contain more pixel-level information but are isolated and noisy. It is therefore difficult to bridge the gap between high- and low-level features due to their difference in terms of physical information content and spatial distribution. In this article, we contribute to solve this problem by enhancing the feature representation in two ways. On the one hand, a patch attention module (PAM) is proposed to enhance the embedding of context information based on a patchwise calculation of local attention. On the other hand, an attention embedding module (AEM) is proposed to enrich the semantic information of low-level features by embedding local focus from high-level features. Both proposed modules are lightweight and can be applied to process the extracted features of convolutional neural networks (CNNs). Experiments show that, by integrating the proposed modules into a baseline fully convolutional network (FCN), the resulting local attention network (LANet) greatly improves the performance over the baseline and outperforms other attention-based methods on two RSI data sets. Numéro de notice : A2021-035 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2020.2994150 Date de publication en ligne : 27/05/2020 En ligne : https://doi.org/10.1109/TGRS.2020.2994150 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=96737
in IEEE Transactions on geoscience and remote sensing > vol 59 n° 1 (January 2021) . - pp 426 - 435[article]Learning-based representations and methods for 3D shape analysis, manipulation and reconstruction / Marie-Julie Rakotosaona (2021)
Titre : Learning-based representations and methods for 3D shape analysis, manipulation and reconstruction Type de document : Thèse/HDR Auteurs : Marie-Julie Rakotosaona, Auteur ; Maks Ovsjanikov, Directeur de thèse Editeur : Palaiseau : Ecole Polytechnique EP Année de publication : 2021 Importance : 148 p. Format : 21 x 30 cm Note générale : bibliographie
These de doctorat de l’Institut Polytechnique de Paris préparée à l’Ecole polytechnique spécialité InformatiqueLangues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image
[Termes IGN] apprentissage profond
[Termes IGN] figure géométrique
[Termes IGN] filtrage de points
[Termes IGN] filtrage du bruit
[Termes IGN] image 3D
[Termes IGN] interpolation
[Termes IGN] maillage
[Termes IGN] maille triangulaire
[Termes IGN] reconstruction 3D
[Termes IGN] semis de points
[Termes IGN] triangulation de Delaunay
[Termes IGN] voxelIndex. décimale : THESE Thèses et HDR Résumé : (auteur) Efficiently processing and analysing 3D data is a crucial challenge in modern applications as 3D shapes are becoming more and more widespread with the proliferation of acquisition devices and modeling tools. While successes of 2D deep learning have become commonplace and surround our daily life, applications that involve 3D data are lagging behind. Due to the more complex non-uniform structure of 3D shapes, successful methods from 2D deep learning cannot be easily extended and there is a strong demand for novel approaches that can both exploit and enable learning using geometric structure. Moreover, being able to handle the various existing representations of 3D shapes such as point clouds and meshes, as well as the artefacts produced from 3D acquisition devices increases the difficulty of the task. In this thesis, we propose systematic approaches that fully exploit geometric information of 3D data in deep learning architectures. We contribute to point cloud denoising, shape interpolation and shape reconstruction methods. We observe that deep learning architectures facilitate learning the underlying surface structure on point clouds that can then be used for denoising as well as shape interpolation. Encoding local patch-based learned priors, as well as complementary geometric information such as edge lengths, leads to powerful pipelines that generate realistic shapes. The key common thread throughout our contributions is facilitating seamless conversion between different representations of shapes. In particular, while using deep learning on triangle meshes is highly challenging due to their combinatorial nature we introduce methods inspired from geometry processing that enable the creation and manipulation of triangle faces. Our methods are robust and generalize well to unseen data despite limited training sets. Our work, therefore, paves the way towards more general, robust and universally useful manipulation of 3D data. Note de contenu : 1- Introduction
2- Introduction en français
3- PointCleanNet: Learning to denoise and remove outliers from dense point clouds
4- Intrinsic point cloud interpolation via dual latent space navigation
5- Learning Delaunay surface elements for mesh reconstruction
6- Differentiable surface triangulation
7- ConclusionNuméro de notice : 28649 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse française Note de thèse : Thèse de Doctorat : Informatique : Ecole Polytechnique : 2021 Organisme de stage : Laboratoire d'informatique de l'École polytechnique DOI : sans En ligne : https://tel.hal.science/tel-03541331/ Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99744
Titre : Learning digital geographies through geographical artificial intelligence Type de document : Thèse/HDR Auteurs : Pengyuan Liu, Auteur ; Stefano de Sabbata, Directeur de thèse ; Yu-Dong Zhang, Directeur de thèse Editeur : Leicester [Royaume-Uni] : University of Leicester Année de publication : 2021 Importance : 199 p. Format : 21 x 30 cm Note générale : bibliographie
A thesis submitted in fulfillment of the requirements for the degree of Doctor of Philosophy, Geology and EnvironmentLangues : Anglais (eng) Descripteur : [Vedettes matières IGN] Analyse spatiale
[Termes IGN] analyse de groupement
[Termes IGN] analyse socio-économique
[Termes IGN] apprentissage profond
[Termes IGN] contenu généré par les utilisateurs
[Termes IGN] croissance urbaine
[Termes IGN] détection de changement
[Termes IGN] données issues des réseaux sociaux
[Termes IGN] données localisées des bénévoles
[Termes IGN] données spatiotemporelles
[Termes IGN] géomatique web
[Termes IGN] intelligence artificielle
[Termes IGN] Londres
[Termes IGN] réseau neuronal de graphes
[Termes IGN] réseau sémantique
[Termes IGN] système d'information urbain
[Termes IGN] zone urbaineIndex. décimale : THESE Thèses et HDR Résumé : (auteur) As the distinction between online and physical spaces rapidly degrades, digital platforms have become an integral component of how people’s everyday experiences are mediated. User-generated content (UGC) shared on such platforms provides insights into how users want to represent their everyday lives, which augments and reinforces our understanding of local communities through time and layers dynamic information across and over the geographic space. Inspired by the development of the newly arisen scientific disciplines within geography: geographical artificial intelligence (GeoAI), this thesis adopts deep learning approaches on graph representations of human dynamics illustrated through geotagged UGC to explore how place representations are augmented and reinforced through users’ spatial experiences by classifying their multimedia activities and identifying the spatial clusters of UGC at the urban scale. Having the place representations described through UGC, this thesis explores how these representations can be used in conjunction with various official spatial statistics to understand and predict the dynamic changes of the socio-economic characteristics of places. The principal contributions of this thesis are: (1) to provide frameworks with higher classification and prediction accuracy but requiring fewer sample data; thus, contributing to an advanced framework to summarise spatial characteristics of places; (2) to show that multimedia content provides rich information regarding places, the use of space, and people’s experience of the landscape; thus, benefiting a better understanding of place representations; (3) to illustrate that the spatial patterns of UGC can be adopted as a valuable proxy to understand urban development and neighbourhood change; (4) to reinforce the concept that Spatial is Special. Spatial processes are commonly spatially autocorrelated. The mainstream of machine learning methods do not explicitly incorporate the spatial or spatio-temporal component to address such a speciality of spatial data. This thesis highlights the importance of explicitly incorporating spatial or spatio-temporal components in geographical analysis models. Note de contenu : 1- Introduction
2- Towards quantitative digital geographies: Concepts, research and implications
3- Data and methods
4- Classification learning through a graph-based semi-supervised approach
5- Location estimation of social media content through a graph-based linkPrediction
6- Urban change modelling with spatial knowledge graphs
7- DiscussionNuméro de notice : 28629 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE/INFORMATIQUE Nature : Thèse étrangère Note de thèse : PhD Thesis: Geology and Environment: Leicester : 2021 DOI : sans En ligne : https://leicester.figshare.com/articles/thesis/Learning_Digital_Geographies_thro [...] Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99618 Learning disentangled representations of satellite image time series in a weakly supervised manner / Eduardo Hugo Sanchez (2021)
Titre : Learning disentangled representations of satellite image time series in a weakly supervised manner Type de document : Thèse/HDR Auteurs : Eduardo Hugo Sanchez, Auteur ; Mathieu Serrurier, Directeur de thèse ; Mathias Ortner, Directeur de thèse Editeur : Toulouse : Université de Toulouse 3 Paul Sabatier Année de publication : 2021 Importance : 176 p. Format : 21 x 30 cm Note générale : bibliographie
Thèse en vue de l'obtention du Doctorat de l'Université de Toulouse, Spécialité Informatique et TélécommunicationsLangues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] analyse des mélanges spectraux
[Termes IGN] analyse des mélanges temporels
[Termes IGN] apprentissage automatique
[Termes IGN] classification dirigée
[Termes IGN] classification non dirigée
[Termes IGN] image Sentinel-MSI
[Termes IGN] réseau antagoniste génératif
[Termes IGN] segmentation d'image
[Termes IGN] série temporelleIndex. décimale : THESE Thèses et HDR Résumé : (auteur) This work focuses on learning data representations of satellite image time series via an unsupervised learning approach. The main goal is to enforce the data representation to capture the relevant information from the time series to perform other applications of satellite imagery. However, extracting information from satellite data involves many challenges since models need to deal with massive amounts of images provided by Earth observation satellites. Additionally, it is impossible for human operators to label such amount of images manually for each individual task (e.g. classification, segmentation, change detection, etc.). Therefore, we cannot use the supervised learning framework which achieves state-of-the-art results in many tasks.To address this problem, unsupervised learning algorithms have been proposed to learn the data structure instead of performing a specific task. Unsupervised learning is a powerful approach since no labels are required during training and the knowledge acquired can be transferred to other tasks enabling faster learning with few labels.In this work, we investigate the problem of learning disentangled representations of satellite image time series where a shared representation captures the spatial information across the images of the time series and an exclusive representation captures the temporal information which is specific to each image. We present the benefits of disentangling the spatio-temporal information of time series, e.g. the spatial information is useful to perform time-invariant image classification or segmentation while the knowledge about the temporal information is useful for change detection. To accomplish this, we analyze some of the most prevalent unsupervised learning models such as the variational autoencoder (VAE) and the generative adversarial networks (GANs) as well as the extensions of these models to perform representation disentanglement. Encouraged by the successful results achieved by generative and reconstructive models, we propose a novel framework to learn spatio-temporal representations of satellite data. We prove that the learned disentangled representations can be used to perform several computer vision tasks such as classification, segmentation, information retrieval and change detection outperforming other state-of-the-art models. Nevertheless, our experiments suggest that generative and reconstructive models present some drawbacks related to the dimensionality of the data representation, architecture complexity and the lack of disentanglement guarantees. In order to overcome these limitations, we explore a recent method based on mutual information estimation and maximization for representation learning without relying on image reconstruction or image generation. We propose a new model that extends the mutual information maximization principle to disentangle the representation domain into two parts. In addition to the experiments performed on satellite data, we show that our model is able to deal with different kinds of datasets outperforming the state-of-the-art methods based on GANs and VAEs. Furthermore, we show that our mutual information based model is less computationally demanding yet more effective. Finally, we show that our model is useful to create a data representation that only captures the class information between two images belonging to the same category. Disentangling the class or category of an image from other factors of variation provides a powerful tool to compute the similarity between pixels and perform image segmentation in a weakly-supervised manner. Note de contenu : Introduction
1- Background
2- Representation disentanglement via VAEs/GANs
3- Representation disentanglement via mutual information estimation
ConclusionNuméro de notice : 24065 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse française Note de thèse : Thèse de Doctorat : Informatique et Télécommunications : Toulouse 3 : 2021 Organisme de stage : nstitut de Recherche en Informatique de Toulouse IRIT DOI : sans En ligne : http://thesesups.ups-tlse.fr/4971/1/2021TOU30032.pdf Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101822 Learning embeddings for cross-time geographic areas represented as graphs / Margarita Khokhlova (2021)
Titre : Learning embeddings for cross-time geographic areas represented as graphs Type de document : Article/Communication Auteurs : Margarita Khokhlova , Auteur ; Nathalie Abadie , Auteur ; Valérie Gouet-Brunet , Auteur ; Liming Chen, Auteur Editeur : New York [Etats-Unis] : Association for computing machinery ACM Année de publication : 2021 Projets : Alegoria / Gouet-Brunet, Valérie Conférence : SAC 2021, 36th Annual ACM Symposium on Applied Computing 22/03/2021 26/03/2021 en ligne Proceedings ACM Importance : pp 559 - 568 Format : 21 x 30 cm Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] arête
[Termes IGN] classification par réseau neuronal
[Termes IGN] entité géographique
[Termes IGN] graphe flou
[Termes IGN] image aérienne à axe vertical
[Termes IGN] noeud
[Termes IGN] relation spatiale
[Termes IGN] représentation graphique
[Termes IGN] réseau neuronal de graphesRésumé : (auteur) Geographic entities from the vertical aerial images can be viewed as discrete objects and represented as nodes in a graph, linked to each other by edges capturing their spatial relationships. Over time, the natural and man made landscape may evolve and thus also their graph representations. This paper addresses the challenging problem of the retrieval and fuzzy matching of graphs to localize near-identical geographical areas across time. Several use-case scenarios are proposed for the end-to-end learning of a graph embedding using Graph Neural Networks (GNN), along with an effective baseline without learning. The results demonstrate the efficiency of our approach, that enables efficient similarity reasoning for novel hand-engineered cross-time graph data. Code and data processing scripts are available online. Numéro de notice : C2021-002 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Thématique : IMAGERIE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.1145/3412841.3441936 En ligne : https://doi.org/10.1145/3412841.3441936 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97583 PermalinkLearning to translate land-cover maps: Several multi-dimensional context-wise solutions / Luc Baudoux (2021)PermalinkLeveraging class hierarchies with metric-guided prototype learning / Vivien Sainte Fare Garnot (2021)PermalinkPermalinkMachine learning for the distributed and dynamic management of a fleet of taxis and autonomous shuttles / Tatiana Babicheva (2021)PermalinkMask R-CNN and OBIA fusion improves the segmentation of scattered vegetation in very high-resolution optical sensors / Emilio Guirado in Sensors, vol 21 n° 1 (January 2021)PermalinkPermalinkA method of hydrographic survey technology selection based on the decision tree supervised learning / Ivana Golub Medvešek (2021)PermalinkPermalinkPermalinkPermalinkPermalinkPermalinkA new method for improving the performance of an ionospheric model developed by multi-instrument measurements based on artificial neural network / Wang Li in Advances in space research, vol 67 n° 1 (January 2021)PermalinkPermalinkPanoptic segmentation of satellite image time series with convolutional temporal attention networks / Vivien Sainte Fare Garnot (2021)PermalinkPermalinkProduction et mise à jour d’un produit BD Forêt V3 par apprentissage profond / Sébastien Giordano (2021)PermalinkPermalinkPermalink