Descripteur
Documents disponibles dans cette catégorie (48)
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
Interactions between hierarchical learning and visual system modeling : image classification on small datasets / Thalita Firmo Drumond (2020)
Titre : Interactions between hierarchical learning and visual system modeling : image classification on small datasets Type de document : Thèse/HDR Auteurs : Thalita Firmo Drumond, Auteur ; Frédéric Alexandre, Directeur de thèse ; Thierry Viéville, Directeur de thèse Editeur : Bordeaux : Université de Bordeaux Année de publication : 2020 Importance : 195 p. Format : 21 x 30 cm Note générale : bibliographie
Thèse présentée pour obtenir le grade de Docteur de l'Université de Bordeaux, Spécialité InformatiqueLangues : Français (fre) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] apprentissage profond
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] classification par réseau neuronal récurrent
[Termes IGN] classification semi-dirigée
[Termes IGN] corpus
[Termes IGN] échantillonnage de données
[Termes IGN] processus de hiérarchisation analytique
[Termes IGN] reconnaissance d'objets
[Termes IGN] taille du jeu de données
[Termes IGN] vision par ordinateurIndex. décimale : THESE Thèses et HDR Résumé : (auteur) Deep convolutional neural networks (DCNN) have recently protagonized a revolution in large-scale object recognition. They have changed the usual computer vision practices of hand-engineered features, with their ability to hierarchically learn representative features from data with a pertinent classifier. Together with hardware advances, they have made it possible to effectively exploit the ever-growing amounts of image data gathered online. However, in specific domains like healthcare and industrial applications, data is much less abundant, and expert labeling costs higher than those of general purpose image datasets. This scarcity scenario leads to this thesis' core question: can these limited-data domains profit from the advantages of DCNNs for image classification? This question has been addressed throughout this work, based on an extensive study of literature, divided in two main parts, followed by proposal of original models and mechanisms.The first part reviews object recognition from an interdisciplinary double-viewpoint. First, it resorts to understanding the function of vision from a biological stance, comparing and contrasting to DCNN models in terms of structure, function and capabilities. Second, a state-of-the-art review is established aiming to identify the main architectural categories and innovations in modern day DCNNs. This interdisciplinary basis fosters the identification of potential mechanisms - inspired both from biological and artificial structures — that could improve image recognition under difficult situations. Recurrent processing is a clear example: while not completely absent from the "deep vision" literature, it has mostly been applied to videos — due to their inherently sequential nature. From biology however it is clear such processing plays a role in refining our perception of a still scene. This theme is further explored through a dedicated literature review focused on recurrent convolutional architectures used in image classification.The second part carries on in the spirit of improving DCNNs, this time focusing more specifically on our central question: deep learning over small datasets. First, the work proposes a more detailed and precise discussion of the small sample problem and its relation to learning hierarchical features with deep models. This discussion is followed up by a structured view of the field, organizing and discussing the different possible paths towards adapting deep models to limited data settings. Rather than a raw listing, this review work aims to make sense out of the myriad of approaches in the field, grouping methods with similar intent or mechanism of action, in order to guide the development of custom solutions for small-data applications. Second, this study is complemented by an experimental analysis, exploring small data learning with the proposition of original models and mechanisms (previously published as a journal paper).In conclusion, it is possible to apply deep learning to small datasets and obtain good results, if done in a thoughtful fashion. On the data path, one shall try gather more information from additional related data sources if available. On the complexity path, architecture and training methods can be calibrated in order to profit the most from any available domain-specific side-information. Proposals concerning both of these paths get discussed in detail throughout this document. Overall, while there are multiple ways of reducing the complexity of deep learning with small data samples, there is no universal solution. Each method has its own drawbacks and practical difficulties and needs to be tailored specifically to the target perceptual task at hand. Note de contenu : 1- Introduction
I- Object recognition with deep convolutional neural networks
2- Convolutional neural networks and visual system modeling
3- Feedforward CNN architectures for object recognition
4- Recurrent and feedback CNN architectures for object recognition
II- Image classification on small datasets
5- A review of strategies to use deep learning under limited data
6- Analysis of DCNN applied to small sample learning using data prototypes
7 ConclusionNuméro de notice : 28312 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse française Note de thèse : thèse de Doctorat : Informatique : Bordeaux : 2020 Organisme de stage : Laboratoire bordelais de recherche en informatique DOI : sans En ligne : https://tel.archives-ouvertes.fr/tel-03129189v2/document Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98233 Optimizing arbovirus surveillance using risk mapping and coverage modelling / Joni A. Downs in Annals of GIS, Vol 26 n° 1 (January 2020)
[article]
Titre : Optimizing arbovirus surveillance using risk mapping and coverage modelling Type de document : Article/Communication Auteurs : Joni A. Downs, Auteur ; Mehrdad Vaziri, Auteur ; George Deskins, Auteur ; et al., Auteur Année de publication : 2020 Article en page(s) : pp 13 - 23 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Analyse spatiale
[Termes IGN] cartographie des risques
[Termes IGN] données environnementales
[Termes IGN] échantillonnage de données
[Termes IGN] épizootie
[Termes IGN] Floride (Etats-Unis)
[Termes IGN] maladie infectieuse
[Termes IGN] modélisation spatiale
[Termes IGN] optimisation spatiale
[Termes IGN] surveillance sanitaire
[Termes IGN] système d'information géographiqueRésumé : (auteur) Diseases carried by mosquitoes and other arthropods endanger human health globally. Though costly, surveillance efforts are vital for disease control and prevention This paper describes an approach for strategically configuring targeted disease surveillance sites across a study area. The methodology combines risk index mapping and spatial optimization modelling. The risk index is used to identify demand for surveillance, and the maximum covering location problem is used to select a specified number of candidate surveillance sites that covers the maximum amount of risk. The approach is demonstrated using a case study where optimal locations for sentinel surveillance sites are selected for the purposes of detecting eastern equine encephalitis virus in a county in the state of Florida. Optimal sentinel sites were selected under a number of scenarios that modelled different target populations (horses or humans), coverage distances (0.5, 1.0, and 1.5 km), and numbers of sites to select (1–12). Sentinel site selections for the horse and human models displayed different spatial patterns, with horse sites located largely in the west-central region and human ones in the north-central. Minor amounts of spatial overlap between the horse and human sites were observed, especially as coverage distances and numbers of sites were increased. Additionally, a near linear increase in risk coverage was observed as sites were incrementally added to the scenarios. This finding suggests that the number of sentinel sites within the ranges explored should be based on the maximum that can be funded, since they provide similar levels of benefit. Numéro de notice : A2020-117 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/19475683.2019.1688391 Date de publication en ligne : 18/11/2019 En ligne : https://doi.org/10.1080/19475683.2019.1688391 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94736
in Annals of GIS > Vol 26 n° 1 (January 2020) . - pp 13 - 23[article]Half 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)
[article]
Titre : Half a percent of labels is enough: efficient animal detection in UAV imagery using deep CNNs and active learning Type de document : Article/Communication Auteurs : Benjamin Kellenberger, Auteur ; Diego Marcos, Auteur ; Sylvain Lobry, Auteur ; Devis Tuia, Auteur Année de publication : 2019 Article en page(s) : pp 9524 - 9533 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] analyse d'image orientée objet
[Termes IGN] apprentissage profond
[Termes IGN] classification orientée objet
[Termes IGN] classification par réseau neuronal
[Termes IGN] détection d'objet
[Termes IGN] données localisées
[Termes IGN] échantillonnage de données
[Termes IGN] faune locale
[Termes IGN] image captée par drone
[Termes IGN] Namibie
[Termes IGN] objet mobile
[Termes IGN] réalité de terrain
[Termes IGN] recensementRésumé : (auteur) We present an Active Learning (AL) strategy for reusing a deep Convolutional Neural Network (CNN)-based object detector on a new data set. This is of particular interest for wildlife conservation: given a set of images acquired with an Unmanned Aerial Vehicle (UAV) and manually labeled ground truth, our goal is to train an animal detector that can be reused for repeated acquisitions, e.g., in follow-up years. Domain shifts between data sets typically prevent such a direct model application. We thus propose to bridge this gap using AL and introduce a new criterion called Transfer Sampling (TS). TS uses Optimal Transport (OT) to find corresponding regions between the source and the target data sets in the space of CNN activations. The CNN scores in the source data set are used to rank the samples according to their likelihood of being animals, and this ranking is transferred to the target data set. Unlike conventional AL criteria that exploit model uncertainty, TS focuses on very confident samples, thus allowing quick retrieval of true positives in the target data set, where positives are typically extremely rare and difficult to find by visual inspection. We extend TS with a new window cropping strategy that further accelerates sample retrieval. Our experiments show that with both strategies combined, less than half a percent of oracle-provided labels are enough to find almost 80% of the animals in challenging sets of UAV images, beating all baselines by a margin. Numéro de notice : A2019-598 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2019.2927393 Date de publication en ligne : 20/08/2019 En ligne : http://doi.org/10.1109/TGRS.2019.2927393 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94592
in IEEE Transactions on geoscience and remote sensing > vol 57 n° 12 (December 2019) . - pp 9524 - 9533[article]A time‐geographic approach to quantifying wildlife–road interactions / Rebecca W. Loraamm in Transactions in GIS, vol 23 n° 1 (February 2019)
[article]
Titre : A time‐geographic approach to quantifying wildlife–road interactions Type de document : Article/Communication Auteurs : Rebecca W. Loraamm, Auteur ; Joni A. Downs, Auteur ; David Lamb, Auteur Année de publication : 2019 Article en page(s) : pp 70 - 86 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Analyse spatiale
[Termes IGN] analyse spatio-temporelle
[Termes IGN] comportement
[Termes IGN] données GPS
[Termes IGN] échantillonnage de données
[Termes IGN] faune locale
[Termes IGN] gestion de la vie sauvage
[Termes IGN] interaction spatiale
[Termes IGN] migration animale
[Termes IGN] Mustelidae
[Termes IGN] prisme spatio-temporel
[Termes IGN] route
[Termes IGN] Time-geography
[Termes IGN] voxelRésumé : (auteur) Recent advances in time geography offer new perspectives for studying animal movements and interactions in an environmental context. In particular, the ability to estimate an animal's spatial location probabilistically at temporal sampling intervals between known fix locations allows researchers to quantify how individuals interact with one another and their environment on finer temporal and spatial scales than previously explored. This article extends methods from time geography, specifically probabilistic space–time prisms, to quantify and summarize animal–road interactions toward understanding related diurnal movement behaviors, including road avoidance. The approach is demonstrated using tracking data for fishers (Martes pennanti) in New York State, where the total probability of interaction with roadways is calculated for individuals over the duration tracked. Additionally, a summarization method visualizing daily interaction probabilities at 60 s intervals is developed to assist in the examination of temporal patterns associated with fishers’ movement behavior with respect to roadways. The results identify spatial and temporal patterns of fisher–roadway interaction by time of day. Overall, the methodologies discussed offer an intuitive means to assess moving object location probabilities in the context of environmental factors. Implications for movement ecology and related conservation planning efforts are also discussed. Numéro de notice : A2019-090 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1111/tgis.12497 Date de publication en ligne : 17/12/2018 En ligne : https://doi.org/10.1111/tgis.12497 Format de la ressource électronique : url Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=92237
in Transactions in GIS > vol 23 n° 1 (February 2019) . - pp 70 - 86[article]Super-resolution of Sentinel-2 images : Learning a globally applicable deep neural network / Charis Lanaras in ISPRS Journal of photogrammetry and remote sensing, vol 146 (December 2018)
[article]
Titre : Super-resolution of Sentinel-2 images : Learning a globally applicable deep neural network Type de document : Article/Communication Auteurs : Charis Lanaras, Auteur ; José Bioucas-Dias, Auteur ; Silvano Galliani, Auteur ; Emmanuel P. Baltsavias, Auteur ; Konrad Schindler, Auteur Année de publication : 2018 Article en page(s) : pp 305 - 319 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] apprentissage profond
[Termes IGN] bande spectrale
[Termes IGN] échantillonnage de données
[Termes IGN] erreur moyenne quadratique
[Termes IGN] image à basse résolution
[Termes IGN] image Sentinel-MSI
[Termes IGN] pansharpening (fusion d'images)
[Termes IGN] pas d'échantillonnage au sol
[Termes IGN] pouvoir de résolution spectrale
[Termes IGN] réseau neuronal convolutifRésumé : (Auteur) The Sentinel-2 satellite mission delivers multi-spectral imagery with 13 spectral bands, acquired at three different spatial resolutions. The aim of this research is to super-resolve the lower-resolution (20 m and 60 m Ground Sampling Distance – GSD) bands to 10 m GSD, so as to obtain a complete data cube at the maximal sensor resolution. We employ a state-of-the-art convolutional neural network (CNN) to perform end-to-end upsampling, which is trained with data at lower resolution, i.e., from 40 20 m, respectively 360 60 m GSD. In this way, one has access to a virtually infinite amount of training data, by downsampling real Sentinel-2 images. We use data sampled globally over a wide range of geographical locations, to obtain a network that generalises across different climate zones and land-cover types, and can super-resolve arbitrary Sentinel-2 images without the need of retraining. In quantitative evaluations (at lower scale, where ground truth is available), our network, which we call DSen2, outperforms the best competing approach by almost 50% in RMSE, while better preserving the spectral characteristics. It also delivers visually convincing results at the full 10 m GSD. Numéro de notice : A2018-540 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2018.09.018 Date de publication en ligne : 21/10/2018 En ligne : https://doi.org/10.1016/j.isprsjprs.2018.09.018 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=91554
in ISPRS Journal of photogrammetry and remote sensing > vol 146 (December 2018) . - pp 305 - 319[article]Réservation
Réserver ce documentExemplaires(3)
Code-barres Cote Support Localisation Section Disponibilité 081-2018131 RAB Revue Centre de documentation En réserve L003 Disponible 081-2018133 DEP-EXM Revue LASTIG Dépôt en unité Exclu du prêt 081-2018132 DEP-EAF Revue Nancy Dépôt en unité Exclu du prêt A review of accuracy assesment for object-based image analysis: from per pixel to per-polygon approaches [review article] / Su Ye in ISPRS Journal of photogrammetry and remote sensing, vol 141 (July 2018)PermalinkSpatially sensitive statistical shape analysis for pedestrian recognition from LIDAR data / Michalis A. Savelonas in Computer Vision and image understanding, vol 171 (June 2018)PermalinkPermalinkGeometric features and their relevance for 3D point cloud classification / Martin Weinmann in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol IV-1/W1 (May 2017)PermalinkUtilisation d’un modèle numérique de hauteur en stratification des données de l’Inventaire Forestier National / Sophie Georges (2017)PermalinkLand-surface segmentation as a method to create strata for spatial sampling and its potential for digital soil mapping / L. Drăguț in International journal of geographical information science IJGIS, vol 30 n° 7- 8 (July - August 2016)PermalinkAn impressionistic cartographic solution for base map land cover with coarse pixel data / Paulo Raposo in Cartographic perspectives, n° 83 (2016)PermalinkUse of remotely sensed auxiliary data for improving sample-based forest inventories / Svetlana Saarela (2015)PermalinkPermalinkMoving and calling: Mobile phone data quality measurements and spatiotemporal uncertainty in human mobility studies / Corina Iovan (2013)Permalink