Centre de documentation
scientifique

Improving deep learning on point cloud by maximizing mutual information across layers / Di Wang in Pattern recognition, vol 131 (November 2022)  

Public [article]  inPattern recognition > vol 131 (November 2022) . - n° 108892 Titre : Improving deep learning on point cloud by maximizing mutual information across layers Type de document : Article/Communication Auteurs : Di Wang, Auteur ; Lulu Tang, Auteur ; Xu Wang, Auteur ; et al., Auteur Année de publication : 2022 Article en page(s) : n° 108892 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] apprentissage profond [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] détection d'objet [Termes IGN] entropie de Shannon [Termes IGN] information sémantique [Termes IGN] segmentation sémantique [Termes IGN] semis de points [Termes IGN] transformation géométrique [Termes IGN] vision par ordinateur [Termes IGN] visualisation 3D Résumé : (auteur) It is a fundamental and vital task to enhance the perception capability of the point cloud learning network in 3D machine vision applications. Most existing methods utilize feature fusion and geometric transformation to improve point cloud learning without paying enough attention to mining further intrinsic information across multiple network layers. Motivated to improve consistency between hierarchical features and strengthen the perception capability of the point cloud network, we propose exploring whether maximizing the mutual information (MI) across shallow and deep layers is beneficial to improve representation learning on point clouds. A novel design of Maximizing Mutual Information (MMI) Module is proposed, which assists the training process of the main network to capture discriminative features of the input point clouds. Specifically, the MMI-based loss function is employed to constrain the differences of semantic information in two hierarchical features extracted from the shallow and deep layers of the network. Extensive experiments show that our method is generally applicable to point cloud tasks, including classification, shape retrieval, indoor scene segmentation, 3D object detection, and completion, and illustrate the efficacy of our proposed method and its advantages over existing ones. Our source code is available at https://github.com/wendydidi/MMI.git. Numéro de notice : A2022-780 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : https://doi.org/10.1016/j.patcog.2022.108892 Date de publication en ligne : 08/07/2022 En ligne : https://doi.org/10.1016/j.patcog.2022.108892 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101859 [article]

The impact of landscape characteristics on the performance of upscaled maps / Peijun Sun in Geocarto international, vol 36 n° 17 ([15/09/2021])  

Public [article]  inGeocarto international > vol 36 n° 17 [15/09/2021] . - pp 1905 - 1922 Titre : The impact of landscape characteristics on the performance of upscaled maps Type de document : Article/Communication Auteurs : Peijun Sun, Auteur ; Russell G. Congalton, Auteur Année de publication : 2021 Article en page(s) : pp 1905 - 1922 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Cartographie [Termes IGN] analyse du paysage [Termes IGN] carte de base [Termes IGN] logiciel de simulation [Termes IGN] mise à l'échelle [Termes IGN] paysage [Termes IGN] précision cartographique Résumé : (auteur) Upscaled maps, as necessary data sources, have drawn much attention to fill data gaps or match the spatial resolution of pre-existing projects. Nevertheless, it remains a challenging task to quantitatively assess the impact of landscape characteristics on the upscaled maps. To simplify the investigation, three characteristics: fragmentation, number of classes and major class impact factor (MCIF), were selected. We utilized SIMMAP to produce categorical maps for generating base maps with different landscape characteristics. The Majority Rule Based algorithm was then used to produce upscaled maps at 11 different spatial resolutions. The findings indicate that the combined effect of landscape patterns greatly impacts upscaling accuracy. This important result should be carefully considered when developing the next generation of upscaling techniques. Overall, extending our understanding of the impacts of landscape characteristics is a critical step forward improving upscaling accuracy and therefore, our use of these maps. Numéro de notice : A2021-693 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/10106049.2019.1678681 Date de publication en ligne : 18/10/2019 En ligne : https://doi.org/10.1080/10106049.2019.1678681 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98511 [article]

Characteristic scales, scaling, and geospatial analysis / Yanguang Chen in Cartographica, vol 56 n° 2 (Summer 2021)  

Public [article]  inCartographica > vol 56 n° 2 (Summer 2021) . - pp 91 -105 Titre : Characteristic scales, scaling, and geospatial analysis Type de document : Article/Communication Auteurs : Yanguang Chen, Auteur Année de publication : 2021 Article en page(s) : pp 91 -105 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Analyse spatiale [Termes IGN] autocorrélation [Termes IGN] dimension fractale [Termes IGN] échelle géographique [Termes IGN] mise à l'échelle [Termes IGN] modèle mathématique Résumé : (auteur) Geographical phenomena fall into two categories: scaleful phenomena and scale-free phenomena. The former have characteristic scales, and the latter have no characteristic scale. Conventional quantitative and mathematical methods can only be applied effectively to scaleful geographical phenomena. In this article, a comparison between scaleful and scale-free geographical systems is drawn by means of simple geographical mathematical models. The main viewpoints are as below. First, scaleful phenomena can be researched by conventional mathematical methods, while scale-free phenomena should be studied using a theory based on scaling such as fractal geometry. Second, the scaleful phenomena belong to distance-based geo-space, while the scale-free phenomena belong to dimension-based geo-space. Third, four approaches to distinguish scale-free phenomena from scaleful phenomena are presented: scaling transform, probability distribution, autocorrelation and partial autocorrelation functions, and ht-index. In practice, a complex geographical system usually possesses scaleful aspects and scale-free aspects. Different methodologies must be adopted for different types of geographic systems or different aspects of the same geographic system. Numéro de notice : A2021-703 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.3138/cart-2020-0001 Date de publication en ligne : 29/05/2021 En ligne : https://doi.org/10.3138/cart-2020-0001 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98592 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
031-2021021	SL	Revue	Centre de documentation	Revues en salle	Disponible

A simplified ICA-based local similarity stereo matching / Suting Chen in The Visual Computer, vol 37 n° 2 (February 2021)  

Public [article]  inThe Visual Computer > vol 37 n° 2 (February 2021) . - pp 411 - 419 Titre : A simplified ICA-based local similarity stereo matching Type de document : Article/Communication Auteurs : Suting Chen, Auteur ; Jinglin Zhang, Auteur ; Meng Jin, Auteur Année de publication : 2021 Article en page(s) : pp 411 - 419 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] analyse en composantes indépendantes [Termes IGN] appariement d'images [Termes IGN] similitude Résumé : (auteur) Since the existing stereo matching methods may fail in the regions of non-textures, boundaries and tiny details, a simplified independent component correlation algorithm (ICA)-based local similarity stereo matching algorithm is proposed. In order to improve the DispNetC, the proposed algorithm first offers the simplified independent component correlation algorithm (SICA) cost aggregation. Then, the algorithm introduces the matching cost volume pyramid, which simplifies the pre-processing process for the ICA. Also, the SICA loss function is defined. Next, the region-wise loss function combined with the pixel-wise loss function is defined as a local similarity loss function to improve the spatial structure of the disparity map. Finally, the SICA loss function is combined with the local similarity loss function, which is defined to estimate the disparity map and to compensate the edge information of the disparity map. Experimental results on KITTI dataset show that the average absolute error of the proposed algorithm is about 37% lower than that of the DispNetC, and its runtime consuming is about 0.6 s lower than that of GC-Net. Numéro de notice : A2021-176 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1007/s00371-020-01811-x Date de publication en ligne : 15/02/2020 En ligne : https://doi.org/10.1007/s00371-020-01811-x Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97286 [article]

BIM/GIS integration for web GIS-based bridge management / Junxiang Zhu in Annals of GIS, vol 27 n° 1 (January 2021)  

Public [article]  inAnnals of GIS > vol 27 n° 1 (January 2021) . - pp 99 - 109 Titre : BIM/GIS integration for web GIS-based bridge management Type de document : Article/Communication Auteurs : Junxiang Zhu, Auteur ; Yi Tan, Auteur ; Wang Xiangyu, Auteur ; Peng Wu, Auteur Année de publication : 2021 Article en page(s) : pp 99 - 109 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications SIG [Termes IGN] ArcGIS online [Termes IGN] Bâti-3D [Termes IGN] CityGML [Termes IGN] format Industry foudation classes IFC [Termes IGN] modélisation 3D du bâti BIM [Termes IGN] pont [Termes IGN] transformation géométrique [Termes IGN] WebSIG Résumé : (auteur) Integration of Building Information Modelling (BIM) and Geographic Information System (GIS) can benefit both the Architecture, Engineering and Construction/Facility Management (AEC/FM) domain, and the geospatial industry. For GIS, BIM is a promising source of three-dimensional (3D) building information that can be used in the study of smart city, green building, big data, etc. The underlying data transformation from Industry Foundation Classes (IFC) to shapefile enables GIS to use the rich building information. However, this transformation is difficult due to the different modelling methods used by IFC and shapefile. IFC generally represents 3D objects using boundary representation (Brep), swept solid, constructive solid geometry (CSG), clipping and mapped representation, while shapefile only uses Brep. In addition, the use of different coordinate systems also contributes to the geometry transformation issue. Among all the solid modelling methods used by IFC, Brep and swept solid are the most often used two methods. Compared with Brep, the transformation of swept solid is more challenging as each type of profile for swept solid defined in IFC has to be properly interpreted and converted. This paper develops algorithms to interpret the commonly used profiles of swept solid for the transformation of IFC into shapefile. A bridge model is used to validate the proposed method, and a Web GIS-based bridge management system is developed to demonstrate a possible usage of the transformed shapefile model. Numéro de notice : A2021-614 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article DOI : 10.1080/19475683.2020.1743355 Date de publication en ligne : 14/04/2020 En ligne : https://doi.org/10.1080/19475683.2020.1743355 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97552 [article]

Contextual land-cover map translation with semantic segmentation / Luc Baudoux (2021)  

Permalink

Developing graphics frameworks with Python and OpenGL / Lee Stemkoski (2021)  

Permalink

Optimisation et développement des solutions photogrammétriques pour la réalisation des relevés de façade au sein du cabinet ELLIPSE Géomètres-Experts / Guillaume Jeannin (2021)  

Permalink

Please, Help Me! I Am Lost in Zoom / Guillaume Touya (2021)  

Permalink

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

Permalink

Context-aware similarity of GPS trajectories / Radu Mariescu-Istodor in Journal of location-based services, vol 14 n° 4 ([01/11/2020])  

Permalink

Thermal unmixing based downscaling for fine resolution diurnal land surface temperature analysis / Jiong Wang in ISPRS Journal of photogrammetry and remote sensing, vol 161 (March 2020)  

Permalink

Learning 3D generation and matching / Thibault Groueix (2020)  

Permalink

A general method for the classification of forest stands using species composition and vertical and horizontal structure / Miquel De Cáceres in Annals of Forest Science, vol 76 n° 2 (June 2019)  

Permalink

A bevy of area-preserving transforms for map projection designers / Daniel "daan" Strebe in Cartography and Geographic Information Science, vol 46 n° 3 (May 2019)  

Permalink