Détail de l'auteur
Auteur Azelle Courtial
|
Documents disponibles écrits par cet auteur (9)
Ajouter le résultat dans votre panier Affiner la recherche Interroger des sources externesConstraint-based evaluation of map images generalized by deep learning / Azelle Courtial in Journal of Geovisualization and Spatial Analysis, vol 6 n° 1 (June 2022)
Titre : Constraint-based evaluation of map images generalized by deep learning Type de document : Article/Communication Auteurs : Azelle Courtial Année de publication : 2022 Projets : 2-Pas d'info accessible - article non ouvert / Article en page(s) : n° 13 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] apprentissage profond
[Termes IGN] connexité (graphes)
[Termes IGN] données d'entrainement (apprentissage automatique)
[Termes IGN] montagne
[Termes IGN] programmation par contraintes
[Termes IGN] qualité des données
[Termes IGN] rendu réaliste
[Termes IGN] route
[Vedettes matières IGN] GénéralisationRésumé : (auteur) Deep learning techniques have recently been experimented for map generalization. Although promising, these experiments raise new problems regarding the evaluation of the output images. Traditional map generalization evaluation cannot directly be applied to the results in a raster format. Additionally, the internal evaluation used by deep learning models is mostly based on the realism of images and the accuracy of pixels, and none of these criteria is sufficient to evaluate a generalization process. Finally, deep learning processes tend to hide the causal mechanisms and do not always guarantee a result that follows cartographic principles. In this article, we propose a method to adapt constraint-based evaluation to the images generated by deep learning models. We focus on the use case of mountain road generalization, and detail seven raster-based constraints, namely, clutter, coalescence reduction, smoothness, position preservation, road connectivity preservation, noise absence, and color realism constraints. These constraints can contribute to current studies on deep learning-based map generalization, as they can help guide the learning process, compare different models, validate these models, and identify remaining problems in the output images. They can also be used to assess the quality of training examples. Numéro de notice : A2022-449 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Thématique : GEOMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1007/s41651-022-00104-2 Date de publication en ligne : 07/05/2022 En ligne : http://dx.doi.org/10.1007/s41651-022-00104-2 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100646
in Journal of Geovisualization and Spatial Analysis > vol 6 n° 1 (June 2022) . - n° 13[article]
Titre : AlpineBends – A benchmark for deep learning-based generalisation Type de document : Article/Communication Auteurs : Azelle Courtial Editeur : ... [Suède] : International Cartographic Association ICA - Association cartographique internationale ACI Année de publication : 2022 Collection : Abstracts of the ICA num. 4 Projets : 1-Pas de projet / Conférence : ICA 2021, 24th ICA Workshop on Map Generalisation and Multiple Representation 13/12/2021 13/12/2021 Florence Italie OA Proceedings Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] apprentissage profond
[Termes IGN] données maillées
[Termes IGN] objet géographique
[Termes IGN] test de performance
[Vedettes matières IGN] GénéralisationRésumé : (auteur) [début] Raster-based map generalization is nowadays anecdotal, as most generalization operations are performed using vector data. Vectors describe the shape of each object in the map using a set of coordinates; thus, the object delimitation is directly accessible, and the topology and distance-based relations are easy to compute. On the contrary, rasters represent a map as an image, a grid of pixel covers the target area, and each pixel is characterised by a value. This representation does not explicitly model the boundary/shape of geographic objects and the relations between them. However, the emergence of the image-based deep learning techniques has shown an ability to process images of geographic information. The question of their adaptation for map generalization is a trendy subject: road (Courtial et al. 2020), building (Feng et al. 2019) and coastline (Du et al. 2021) generalization have been explored in recent years. Common methods for evaluating these techniques seems to be necessary for the comparison and development of this field. Numéro de notice : C2021-067 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Thématique : GEOMATIQUE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.5194/ica-abs-4-1-2022 Date de publication en ligne : 14/01/2022 En ligne : https://doi.org/10.5194/ica-abs-4-1-2022 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99535
Titre : BasqueRoads: a benchmark for road network selection Type de document : Article/Communication Auteurs : Guillaume Touya Editeur : ... [Suède] : International Cartographic Association ICA - Association cartographique internationale ACI Année de publication : 2022 Collection : Abstracts of the ICA num. 4 Projets : LostInZoom / Touya, Guillaume Conférence : ICA 2021, 24th ICA Workshop on Map Generalisation and Multiple Representation 13/12/2021 13/12/2021 Florence Italie OA Proceedings Note générale : bibliographie
This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 101003012).Langues : Anglais (eng) Descripteur : [Termes IGN] objet géographique
[Termes IGN] réseau routier
[Termes IGN] simplification de contour
[Termes IGN] test de performance
[Vedettes matières IGN] GénéralisationRésumé : (auteur) [début] Road network selection is one of the major issues of map generalisation, as new papers are proposed every year since the first attempts of automation in the 1990’s (Thomson & Richardson, 1995). New methods are regularly proposed because selecting roads for maps at smaller scales is a complex problem. Roads are at the same time present in maps to enable car navigation tasks, and because they are structuring elements that reveal the nature of the landscape (urban, rural, mountainous…). So road selection is not only about retaining the most important roads of the network, but the preservation of topology and connectivity is essential, as well as the preservation, or the typification of road patterns (e.g. a ring road), and the preservation of local density differences (between urban and rural areas for instance). It is rare to see comparisons of road selection techniques in the literature, because of the lack of open source in map generalisation, but also because of the lack of a common dataset to benchmark these techniques; new propositions on road selection are most of the time tied to their own dataset and use case. This is why we think that this BasqueRoads dataset could be useful to advance on this topic of road network selection. Numéro de notice : C2021-066 Affiliation des auteurs : UGE-LASTIG (2020- ) Thématique : GEOMATIQUE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.5194/ica-abs-4-5-2022 Date de publication en ligne : 14/01/2022 En ligne : https://doi.org/10.5194/ica-abs-4-5-2022 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99536
Titre : Representing vector geographic information as a tensor for deep learning based map generalisation Type de document : Article/Communication Auteurs : Azelle Courtial Editeur : AGILE Alliance Année de publication : 2022 Conférence : AGILE 2022, 25th international AGILE Conference on Geographic Information Science, Artificial intelligence in the service of geospatial technologies 14/06/2022 17/06/2022 Vilnius Lithuanie OA Proceedings Importance : 8 p. Format : 21 x 30 cm Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] alignement des données
[Termes IGN] apprentissage profond
[Termes IGN] architecture de réseau
[Termes IGN] bati
[Termes IGN] carte topographique
[Termes IGN] couche
[Termes IGN] données d'entrainement (apprentissage automatique)
[Termes IGN] données vectorielles
[Termes IGN] information sémantique
[Termes IGN] milieu urbain
[Termes IGN] route
[Termes IGN] tenseur
[Vedettes matières IGN] GénéralisationRésumé : (auteur) Recently, many researchers tried to generate (generalised) maps using deep learning, and most of the proposed methods deal with deep neural network architecture choices. Deep learning learns to reproduce examples, so we think that improving the training examples, and especially the representation of the initial geographic information, is the key issue for this problem. Our article extracts some representation issues from a literature review and proposes different ways to represent vector geographic information as a tensor. We propose two kinds of contributions: 1) the representation of information by layers; 2) the representation of additional information. Then, we demonstrate the interest of some of our propositions with experiments that show a visual improvement for the generation of generalised topographic maps in urban areas. Numéro de notice : C2022-024 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Autre URL associée : https://agile-giss.copernicus.org/articles/3/index.html Thématique : GEOMATIQUE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.5194/agile-giss-3-32-2022 En ligne : https://doi.org/10.5194/agile-giss-3-32-2022 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100921
Titre : Can graph convolution networks learn spatial relations? Type de document : Article/Communication Auteurs : Azelle Courtial Editeur : ... [Suède] : International Cartographic Association ICA - Association cartographique internationale ACI Année de publication : 2021 Collection : Abstracts of the ICA num. 3 Projets : 1-Pas de projet / Touya, Guillaume Conférence : ICC 2021, 30th ICA international cartographic conference 14/12/2021 18/12/2021 Florence Italie OA Proceedings Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] alignement
[Termes IGN] bati
[Termes IGN] objet géographique
[Termes IGN] relation spatiale
[Termes IGN] réseau neuronal convolutif
[Termes IGN] réseau neuronal de graphes
[Termes IGN] réseau routier
[Vedettes matières IGN] GénéralisationRésumé : (auteur) [introduction] Maps are composed of spatially related geographic objects. Spatial relations are key information for human as they support the description of relative locations: the house is to the east of the city centre, near the interchange, or at the end of the path. Consequently, preserving these spatial relations is important during map generalisation. For example, building typification is a generalisation operation that seeks to reduce the quantity of building while preserving relation between and within homogeneous buildings groups (Regnauld, 2001). Building or road patterns are remarkable distributions of elements in the map from which high-level concepts and semantics (e.g. landuse types and urban morphology) can be inferred. Such patterns can be characterized by spatial relations (e.g. proximity, similarity and continuity of these elements) and hence are visually easy to identify by a human. To identify these patterns automatically is important for automated map generalisation (Christophe and Ruas, 2002). However, it remains challenging to devise algorithms that can resemble the human level performance. The goal of this paper is to illustrate the potential of graph convolutional networks (GCN) for the identification of patterns and relations important for map generalisation with two use cases: building patterns detection, and road segment selection. Both tasks require some degree of understanding of the spatial relations between map objects. Hence, our experiments constitute a first step in exploring the capability of deep neural network for learning representations of spatial relations. Numéro de notice : C2021-045 Affiliation des auteurs : UGE-LASTIG (2020- ) Thématique : GEOMATIQUE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.5194/ica-abs-3-60-2021 Date de publication en ligne : 13/12/2021 En ligne : https://doi.org/10.5194/ica-abs-3-60-2021 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99420 PermalinkPermalinkPermalinkPermalink
