Centre dedocumentation
scientifique

Scalable surface reconstruction with Delaunay-Graph neural networks / Raphaël Sulzer in Computer graphics forum, vol 40 n° 5 (2021)  

Public [article]  inComputer graphics forum > vol 40 n° 5 (2021) . - pp 157 - 167 Titre : Scalable surface reconstruction with Delaunay-Graph neural networks Type de document : Article/Communication Auteurs : Raphaël Sulzer, Auteur ; Loïc Landrieu , Auteur ; Renaud Marlet, Auteur ; Bruno Vallet , Auteur Année de publication : 2021 Projets : BIOM / Vallet, Bruno Article en page(s) : pp 157 - 167 Note générale : bibliographie The presentation of this work at SGP 2021 is available at https://youtu.be/KIrCDGhS10o Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications photogrammétriques [Termes IGN] algorithme Graph-Cut [Termes IGN] apprentissage profond [Termes IGN] prise en compte du contexte [Termes IGN] reconstruction d'objet [Termes IGN] réseau neuronal de graphes [Termes IGN] semis de points [Termes IGN] tétraèdre [Termes IGN] triangulation de Delaunay Résumé : (auteur) We introduce a novel learning-based, visibility-aware, surface reconstruction method for large-scale, defect-laden point clouds. Our approach can cope with the scale and variety of point cloud defects encountered in real-life Multi-View Stereo (MVS) acquisitions. Our method relies on a 3D Delaunay tetrahedralization whose cells are classified as inside or outside the surface by a graph neural network and an energy model solvable with a graph cut. Our model, making use of both local geometric attributes and line-of-sight visibility information, is able to learn a visibility model from a small amount of synthetic training data and generalizes to real-life acquisitions. Combining the efficiency of deep learning methods and the scalability of energy-based models, our approach outperforms both learning and non learning-based reconstruction algorithms on two publicly available reconstruction benchmarks. Numéro de notice : A2021-400 Affiliation des auteurs : UGE-LaSTIG+Ext (2020- ) Thématique : IMAGERIE/INFORMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1111/cgf14364 En ligne : https://doi.org/10.1111/cgf.14364 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98219 [article]

Multi-modal learning in photogrammetry and remote sensing / Michael Ying Yang in ISPRS Journal of photogrammetry and remote sensing, vol 176 (June 2021)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 176 (June 2021) . - pp 54 - 54 Titre : Multi-modal learning in photogrammetry and remote sensing Type de document : Article/Communication Auteurs : Michael Ying Yang, Auteur ; Loïc Landrieu , Auteur ; Devis Tuia, Auteur ; Charles Toth, Auteur Année de publication : 2021 Projets : 1-Pas de projet / Vallet, Bruno Article en page(s) : pp 54 - 54 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Télédétection [Termes IGN] acquisition d'images [Termes IGN] apprentissage automatique [Termes IGN] données multisources Résumé : (Auteur) [Editorial] There is a growing interest in the photogrammetry and remote sensing community for multi-modal data, i. e., data simultaneously acquired from a variety of platforms, including satellites, aircraft, UAS/UGS, autonomous vehicles, etc., by different sensors, such as radar, optical, LiDAR. Thanks to their different spatial, spectral, or temporal resolutions, the use of complementary data sources leads to richer and more robust information extraction. We expect that the use of multiple modalities will rapidly become a standard approach in the future. The main difficulty of jointly processing multi-modal data is due to the differences in structure among modalities. Another issue is the unbalanced number of labelled samples available across modalities, resulting in a significant gap in performance when models are trained separately. Clearly, the photogrammetry and remote sensing community has not yet exploited the full potential of multi-modal data. Neural networks seem well suited for accommodating different data sources, thanks to their capabilities to learn representations adapted to each task in an end-to-end fashion. In this context, there is a strong need for research and development of approaches for multi-sensory and multi-modal deep learning within the geospatial domain. Numéro de notice : A2021-364 Affiliation des auteurs : UGE-LaSTIG+Ext (2020- ) Thématique : IMAGERIE/INFORMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2021.03.022 date de publication en ligne : 23/04/2021 En ligne : https://doi.org/10.1016/j.isprsjprs.2021.03.022 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97660 [article]

Representing shape collections with alignment-aware linear models / Romain Loiseau (2021)  

Public Titre : Representing shape collections with alignment-aware linear models Type de document : Article/Communication Auteurs : Romain Loiseau, Auteur ; Tom Monnier, Auteur ; Loïc Landrieu , Auteur ; Mathieu Aubry, Auteur Editeur : Ithaca [New York - Etats-Unis] : ArXiv - Université Cornell Année de publication : 2021 Projets : READY3D / Landrieu, Loïc Importance : 17 p. Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] analyse de données [Termes IGN] apprentissage profond [Termes IGN] données localisées 3D [Termes IGN] modèle linéaire [Termes IGN] réseau neuronal profond [Termes IGN] segmentation [Termes IGN] semis de points [Termes IGN] transformation affine Résumé : (auteur) In this paper, we revisit the classical representation of 3D point clouds as linear shape models. Our key insight is to leverage deep learning to represent a collection of shapes as affine transformations of low-dimensional linear shape models. Each linear model is characterized by a shape prototype, a low-dimensional shape basis and two neural networks. The networks take as input a point cloud and predict the coordinates of a shape in the linear basis and the affine transformation which best approximate the input. Both linear models and neural networks are learned end-to-end using a single reconstruction loss. The main advantage of our approach is that, in contrast to many recent deep approaches which learn feature-based complex shape representations, our model is explicit and every operation occurs in 3D space. As a result, our linear shape models can be easily visualized and annotated, and failure cases can be visually understood. While our main goal is to introduce a compact and interpretable representation of shape collections, we show it leads to state of the art results for few-shot segmentation. Code and data are available at: https://romainloiseau.github.io/deep-linear-shapes Numéro de notice : P2021-005 Affiliation des auteurs : UGE-LaSTIG+Ext (2020- ) Thématique : INFORMATIQUE Nature : Preprint nature-HAL : Préprint DOI : sans date de publication en ligne : 03/09/2021 En ligne : https://arxiv.org/abs/2109.01605 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98385

Improved crop classification with rotation knowledge using Sentinel-1 and -2 time series / Sébastien Giordano in Photogrammetric Engineering & Remote Sensing, PERS, vol 86 n° 7 (July 2020)  

Public [article]  inPhotogrammetric Engineering & Remote Sensing, PERS > vol 86 n° 7 (July 2020) . - pp 431 - 441 Titre : Improved crop classification with rotation knowledge using Sentinel-1 and -2 time series Type de document : Article/Communication Auteurs : Sébastien Giordano , Auteur ; Simon Bailly, Auteur ; Loïc Landrieu , Auteur ; Nesrine Chehata , Auteur Année de publication : 2020 Projets : MAESTRIA / Mallet, Clément Article en page(s) : pp 431 - 441 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image mixte [Termes IGN] Alpes-de-haute-provence (04) [Termes IGN] chaîne de traitement [Termes IGN] classification par forêts aléatoires [Termes IGN] image Sentinel-MSI [Termes IGN] image Sentinel-SAR [Termes IGN] parcelle agricole [Termes IGN] photo-identification [Termes IGN] Seine-et-Marne (77) [Termes IGN] série temporelle [Termes IGN] surface cultivée Résumé : (Auteur) Leveraging the recent availability of accurate, frequent, and multimodal (radar and optical) Sentinel-1 and -2 acquisitions, this paper investigates the automation of land parcel identification system (LPIS) crop type classification. Our approach allows for the automatic integration of temporal knowledge, i.e., crop rotations using existing parcel-based land cover databases and multi-modal Sentinel-1 and -2 time series. The temporal evolution of crop types was modeled with a linear-chain conditional random field, trained with time series of multimodal (radar and optical) satellite acquisitions and associated LPIS. Our model was tested on two study areas in France (≥ 1250 km2) which show different crop types, various parcel sizes, and agricultural practices: . the Seine et Marne and the Alpes de Haute-Provence classified accordingly to a fine national 25-class nomenclature. We first trained a Random Forest classifier without temporal structure to achieve 89.0% overall accuracy in Seine et Marne (10 classes) and 73% in Alpes de Haute-Provence (14 classes). We then demonstrated experimentally that taking into account the temporal structure of crop rotation with our model resulted in an increase of 3% to +5% in accuracy. This increase was especially important (+12%) for classes which were poorly classified without using the temporal structure. A stark positive impact was also demonstrated on permanent crops, while it was fairly limited or even detrimental for annual crops. Numéro de notice : A2020-382 Affiliation des auteurs : LaSTIG MATIS+Ext (2012-2019) Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.14358/PERS.86.7.431 date de publication en ligne : 01/07/2020 En ligne : https://doi.org/10.14358/PERS.86.7.431 Format de la ressource électronique : URL Article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=95428 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
105-2020071	SL	Revue	Centre de documentation	Revues en salle	Disponible

Lightweight temporal self-attention for classifying satellite images time series / Vivien Sainte Fare Garnot (2020)  

Public Titre : Lightweight temporal self-attention for classifying satellite images time series Type de document : Article/Communication Auteurs : Vivien Sainte Fare Garnot , Auteur ; Loïc Landrieu , Auteur Editeur : Le Chesnay : Institut National de Recherche en Informatique et en Automatique INRIA Année de publication : 2020 Conférence : AALTD 2020, Workshop on Advanced Analytics and Learning on Temporal Data 18/09/2020 18/09/2020 en ligne Belgique Importance : 11 p. Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] chaîne de traitement [Termes IGN] efficacité [Termes IGN] image Sentinel [Termes IGN] série temporelle Résumé : (auteur) The increasing accessibility and precision of Earth observation satellite data offers considerable opportunities for industrial and state actors alike. This calls however for efficient methods able to process time series on a global scale. Building on recent work employing multi-headed self-attention mechanisms to classify remote sensing time sequences, we propose a modification of the Temporal Attention Encoderof Garnot et al.[5]. In our network, the channels of the temporal inputs are distributed among several compact attention heads operating in parallel. Each head extracts highly-specialized temporal features which arein turn concatenated into a single representation. Our approach outper-forms other state-of-the-art time series classification algorithms on an open-access satellite image dataset, while using significantly fewer parameters and with a reduced computational complexity. Numéro de notice : C2020-020 Affiliation des auteurs : UGE-LaSTIG (2020- ) Autre URL associée : vers HAL Thématique : IMAGERIE/INFORMATIQUE Nature : Poster nature-HAL : Poster-avec-CL DOI : sans En ligne : https://project.inria.fr/aaltd20/files/2020/08/AALTD_20_paper_SainteFareGarnot.p [...] Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=96458

Satellite image time series classification with pixel-set encoders and temporal self-attention / Vivien Sainte Fare Garnot (2020)    

Permalink

Torch-Points3D: A modular multi-task framework for reproducible deep learning on 3D point clouds / Thomas Chaton (2020)  

Permalink

Piecewise-planar approximation of large 3D data as graph-structured optimization / Stéphane Guinard in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, IV-2/W5 (May 2019)  

Permalink

Parallel cut pursuit for minimization of the graph total variation / Hugo Raguet (2019)  

Permalink

Point cloud oversegmentation with graph-structured deep metric learning / Loïc Landrieu (2019)  

Permalink

Supervized segmentation with graph-structured deep metric learning / Loïc Landrieu (2019)  

Permalink

Time-space tradeoff in deep learning models for crop classification on satellite multi-spectral image time series / Vivien Sainte Fare Garnot (2019)  

Permalink

Crop-rotation structured classification using multi-source sentinel images and LPIS for crop type mapping / Simon Bailly (2018)  

Permalink

Cut-Pursuit algorithm for regularizing nonsmooth functionals with graph total variation / Hugo Raguet (2018)    

Permalink

Large-scale point cloud semantic segmentation with superpoint graphs / Loïc Landrieu (2018)    

Permalink