Descripteur
Termes IGN > informatique > intelligence artificielle > apprentissage automatique > apprentissage non-dirigé
apprentissage non-dirigéVoir aussi |
Documents disponibles dans cette catégorie (82)
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
Clustering et apprentissage profond sous contraintes pour l’analyse de séries temporelles : Application à l’analyse temporelle incrémentale en télédétection / Baptiste Lafabregue (2021)
Titre : Clustering et apprentissage profond sous contraintes pour l’analyse de séries temporelles : Application à l’analyse temporelle incrémentale en télédétection Type de document : Thèse/HDR Auteurs : Baptiste Lafabregue, Auteur ; Germain Forestier, Directeur de thèse ; Pierre Gançarski, Directeur de thèse Editeur : Mulhouse : Université de Haute Alsace Année de publication : 2021 Importance : 167 p. Format : 21 x 30 cm Note générale : bibliographie
Thèse présentée pour obtenir le grade de Docteur de l'Université de Haute-Alsace, Discipline InformatiqueLangues : Français (fre) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] analyse de groupement
[Termes IGN] analyse spatio-temporelle
[Termes IGN] apprentissage non-dirigé
[Termes IGN] apprentissage profond
[Termes IGN] jeu de données
[Termes IGN] programmation par contraintes
[Termes IGN] segmentation sémantique
[Termes IGN] série temporelleIndex. décimale : THESE Thèses et HDR Résumé : (auteur) Depuis quelques années, les satellites réalisent des captures d'images de la Terre avec une haute fréquence de revisite et une haute disponibilité, qu'on peut représenter sous forme de séries temporelles. Cela permet d'effectuer une observation continue de la Terre avec des applications dans le suivi agricole, la gestion de catastrophes naturelles, etc. Cependant, ce phénomène ne se limite pas au domaine de la télédétection. On peut en effet observer une croissance similaire dans de nombreux domaines, tel que la médecine ou la finance. Or, dans tous ces domaines, l'analyse de ces données fait face aux mêmes problématiques. Une grande quantité de données n'est pas toujours accompagnée d'un étiquetage suffisant, ce qui empêche généralement une bonne application des méthodes supervisées. En effet, l'étiquetage reste une tâche très chronophage et complexe, car nécessitant une expertise sur les données analysées. A l'opposé, les méthodes non supervisées ne nécessitent pas de connaissances de l'expert mais donnent parfois des résultats médiocres. Dans ce contexte, le clustering sous contraintes est une alternative qui offre un bon compromis en termes d'investissement pour l'expert. Toutefois, les méthodes de clustering sous contraintes sont sujettes à des limitations importantes. Nous montrons dans cette thèse que deux facteurs limites fortement l'impact des contraintes, la consistance, qui est la quantité d'information dans l'ensemble des contraintes que l'algorithme peut déterminer par ses propres biais, et la cohérence, qui est le degré d'accord entre les contraintes elles-mêmes. Afin de répondre au problème de consistance, nous proposons une nouvelle méthode, I-SAMARAH, basée sur le clustering collaboratif et l'intégration des contraintes de manière incrémentale. Cependant, nous montrons également que le problème de cohérence reste important que nous proposons d'aborder de manière plus prospective avec des méthodes basées sur l'apprentissage profond. Note de contenu : Introduction
1- Contexte
2- Guider le clustering avec des contraintes
3- Analyse de séries temporelles en télédétection
4- Apprentissage de représentation contraint
5- Apprentissage profond non-supervisé et séries temporelles
ConclusionNuméro de notice : 15276 Affiliation des auteurs : non IGN Thématique : IMAGERIE/INFORMATIQUE Nature : Thèse française Note de thèse : Thèse de Doctorat : Informatique : Haute Alsace : 2021 Organisme de stage : IRIMAS DOI : sans En ligne : https://tel.hal.science/tel-03630122 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101099
Titre : Context-aware image super-resolution using deep neural networks Type de document : Thèse/HDR Auteurs : Mohammad Saeed Rad, Auteur ; Jean-Philippe Thiran, Directeur de thèse Editeur : Lausanne : Ecole Polytechnique Fédérale de Lausanne EPFL Année de publication : 2021 Importance : 148 p. Format : 21 x 30 cm Note générale : bibliographie
Thèse présentée pour l'obtention du grade de Docteur ès SciencesLangues : Français (fre) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] apprentissage profond
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] image à basse résolution
[Termes IGN] image à haute résolution
[Termes IGN] pouvoir de résolution spectrale
[Termes IGN] reconstruction d'image
[Termes IGN] réseau antagoniste génératif
[Termes IGN] segmentation sémantique
[Termes IGN] vision par ordinateurIndex. décimale : THESE Thèses et HDR Résumé : (auteur) Image super-resolution is a classic ill-posed computer vision and image processing problem, addressing the question of how to reconstruct a high-resolution image from its low-resolution counterpart. Current state-of-the-art methods have improved the performance of the single image super-resolution task significantly by benefiting from machine learning and AI-powered algorithms, and more specifically, with the advent of Deep Learning-based approaches. Although these advances allow a machine to learn and have better exploitation of an image and its content, recent methods are still unable to constrain the plausible solution space based on the available contextual information within an image. This limitation mostly results in poor reconstructions, even for well-known types of objects and textures easily recognizable for humans. In this thesis, we aim at proving that the categorical prior, which characterizes the semantic class of a region in an image (e.g., sky, building, plant), is crucial in super-resolution task for reaching a higher reconstruction quality. In particular, we propose several approaches to improve the perceived image quality and generalization capability of deep learning-based methods by exploiting the context and semantic meaning of images. To prove the effectiveness of this categorical information, we first propose a convolutional neural network-based framework that is able to extract and use semantic information to super-resolve a given image by using multitask learning, simultaneously for learning image super-resolution and semantic segmentation. The proposed decoder is forced to explore categorical information during training, as this setting employs only one shared deep network for both semantic segmentation and super-resolution tasks. We further investigate the possibility of using semantic information by a novel objective function to introduce additional spatial control over the training process. We propose penalizing images at different semantic levels using appropriate loss terms by benefiting from our new OBB (Object, Background, and Boundary) labels generated from segmentation labels. Then, we introduce a new test time adaptation-based technique to leverage high-resolution images with perceptually similar context to a given test image to improve the reconstruction quality. We further validate this approach's effectiveness by using a novel numerical experiment analyzing the correlation between filters learned by our network and what we define as `ideal' filters. Finally, we present a generic solution to enable adapting all our previous contributions in this thesis, as well as other recent super-resolution works trained on synthetic datasets, to real-world super-resolution problem. Real-world super-resolution refers to super-resolving images with real degradations caused by physical imaging systems, instead of low-resolution images from simulated datasets assuming a simple and uniform degradation model (i.e., bicubic downsampling). We study and develop an image-to-image translator to map the distribution of real low-resolution images to the well-understood distribution of bicubically downsampled images. This translator is used as a plug-in to integrate real inputs into any super-resolution framework trained on simulated datasets. We carry out extensive qualitative and quantitative experiments for each mentioned contribution, including user studies, to compare our proposed approaches to state-of-the-art method. Note de contenu : 1- Introduction
2- Brief image super-resolution review
3- Extracting image context by multi-task learning
4- Spatial control over image genertion process
5- Test-time adaptation based on perceptual similarity
6- Integrating into real-world SR
7- ConclusionNuméro de notice : 28652 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Note de thèse : Thèse de Doctorat : Sciences : EPFL, Lausanne : 2021 DOI : sans En ligne : https://infoscience.epfl.ch/record/286804?ln=fr Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99790
Titre : Deep-learning for 3D reconstruction Type de document : Thèse/HDR Auteurs : Fabio Tosi, Auteur Editeur : Bologne [Italie] : Université de Bologne Année de publication : 2021 Format : 21 x 30 cm Note générale : bibliographie
PhD Thesis in Computer Science and EngineeringLangues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] apprentissage automatique
[Termes IGN] apprentissage profond
[Termes IGN] carte de confiance
[Termes IGN] compréhension de l'image
[Termes IGN] profondeur
[Termes IGN] reconstruction 3D
[Termes IGN] réseau antagoniste génératif
[Termes IGN] vision stéréoscopiqueRésumé : (auteur) Depth perception is paramount for many computer vision applications such as autonomous driving and augmented reality. Despite active sensors (e.g., LiDAR, Time-of-Flight, struc- tured light) are quite diffused, they have severe shortcomings that could be potentially addressed by image-based sensors. Concerning this latter category, deep learning has enabled ground-breaking results in tackling well-known issues affecting the accuracy of systems inferring depth from a single or multiple images in specific circumstances (e.g., low textured regions, depth discontinuities, etc.), but also introduced additional concerns about the domain shift occurring between training and target environments and the need of proper ground truth depth labels to be used as the training signals in network learning. Moreover, despite the copious literature concerning confidence estimation for depth from a stereo setup, inferring depth uncertainty when dealing with deep networks is still a major challenge and almost unexplored research area, especially when dealing with a monocular setup. Finally, computational complexity is another crucial aspect to be considered when targeting most practical applications and hence is desirable not only to infer reliable depth data but do so in real-time and with low power requirements even on standard embedded devices or smartphones. Therefore, focusing on stereo and monocular setups, this thesis tackles major issues affecting methodologies to infer depth from images and aims at developing accurate and efficient frameworks for accurate 3D reconstruction on challenging environments. Note de contenu : Introduction
1- Related work
2- Datasets
3- Evaluation protocols
4- Confidence measures in a machine learning world
5- Efficient confidence measures for embedded stereo
6- Even more confident predictions with deep machine-learning
7- Beyond local reasoning for stereo confidence estimation with deep learning
8- Good cues to learn from scratch a confidence measure for passive depth sensors
9- Confidence estimation for ToF and stereo sensors and its application to depth data fusion
10- Learning confidence measures in the wild
11- Self-adapting confidence estimation for stereo
12- Leveraging confident points for accurate depth refinement on embedded systems
13- SMD-Nets: Stereo Mixture Density Networks
14- Real-time self-adaptive deep stereo
15- Guided stereo matching
16- Reversing the cycle: self-supervised deep stereo through enhanced monocular distillation
17- Learning end-to-end scene flow by distilling single tasks knowledge
18- Learning monocular depth estimation with unsupervised trinocular assumptions
19- Geometry meets semantics for semi-supervised monocular depth estimation
20- Generative Adversarial Networks for unsupervised monocular depth prediction
21- Learning monocular depth estimation infusing traditional stereo knowled
22- Towards real-time unsupervised monocular depth estimation on CPU
23- Enabling energy-efficient unsupervised monocular depth estimation on ARMv7-based platforms
24- Distilled semantics for comprehensive scene understanding from videos
25- On the uncertainty of self-supervised monocular depth estimation
ConclusionNuméro de notice : 28596 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Note de thèse : Thèse de Doctorat : Computer Science and Engineering : Bologne : 2021 DOI : 10.48676/unibo/amsdottorato/9816 En ligne : http://amsdottorato.unibo.it/9816/ Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99325 Generative adversarial networks to generalise urban areas in topographic maps / Azelle Courtial (2021)
Titre : Generative adversarial networks to generalise urban areas in topographic maps Type de document : Article/Communication Auteurs : Azelle Courtial , Auteur ; Guillaume Touya , Auteur ; Xiang Zhang, Auteur Editeur : International Society for Photogrammetry and Remote Sensing ISPRS Année de publication : 2021 Collection : International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, ISSN 1682-1750 num. 43-B4-2021 Projets : 1-Pas de projet / Conférence : ISPRS 2021, Commission 4, XXIV ISPRS Congress, Imaging today foreseeing tomorrow 05/07/2021 09/07/2021 Nice Virtuel France OA Archives Commission 4 Importance : pp 15 - 22 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] apprentissage profond
[Termes IGN] carte topographique
[Termes IGN] généralisation cartographique automatisée
[Termes IGN] réseau antagoniste génératif
[Termes IGN] zone urbaine
[Vedettes matières IGN] GénéralisationRésumé : (auteur) This article presents how a generative adversarial network (GAN) can be employed to produce a generalised map that combines several cartographic themes in the dense context of urban areas. We use as input detailed buildings, roads, and rivers from topographic datasets produced by the French national mapping agency (IGN), and we expect as output of the GAN a legible map of these elements at a target scale of 1:50,000. This level of detail requires to reduce the amount of information while preserving patterns; covering dense inner cities block by a unique polygon is also necessary because these blocks cannot be represented with enlarged individual buildings. The target map has a style similar to the topographic map produced by IGN. This experiment succeeded in producing image tiles that look like legible maps. It also highlights the impact of data and representation choices on the quality of predicted images, and the challenge of learning geographic relationships. Numéro de notice : C2021-016 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Thématique : GEOMATIQUE/INFORMATIQUE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.5194/isprs-archives-XLIII-B4-2021-15-2021 Date de publication en ligne : 30/06/2021 En ligne : https://doi.org/10.5194/isprs-archives-XLIII-B4-2021-15-2021 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98062 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 Spectral variability in hyperspectral unmixing : Multiscale, tensor, and neural network-based approaches / Ricardo Augusto Borsoi (2021)PermalinkPermalinkPermalinkUnderstanding the role of individual units in a deep neural network / David Bau in Proceedings of the National Academy of Sciences of the United States of America PNAS, vol 117 n° 48 (1 December 2020)PermalinkRiver ice segmentation with deep learning / Abhineet Singh in IEEE Transactions on geoscience and remote sensing, vol 58 n° 11 (November 2020)PermalinkOpenStreetMap quality assessment using unsupervised machine learning methods / Kent T. Jacobs in Transactions in GIS, Vol 24 n° 5 (October 2020)PermalinkPermalinkPermalinkUnsupervised satellite image time series analysis using deep learning techniques / Ekaterina Kalinicheva (2020)PermalinkVideo event recognition and anomaly detection by combining gaussian process and hierarchical dirichlet process models / Michael Ying Yang in Photogrammetric Engineering & Remote Sensing, PERS, vol 84 n° 4 (April 2018)Permalink