Centre de documentation
scientifique

A geometry-aware attention network for semantic segmentation of MLS point clouds / Jie Wan in International journal of geographical information science IJGIS, vol 37 n° 1 (January 2023)  

Public [article]  inInternational journal of geographical information science IJGIS > vol 37 n° 1 (January 2023) . - pp 138 - 161 Titre : A geometry-aware attention network for semantic segmentation of MLS point clouds Type de document : Article/Communication Auteurs : Jie Wan, Auteur ; Yongyang Xu, Auteur ; Qinjun Qiu, Auteur ; et al., Auteur Année de publication : 2023 Article en page(s) : pp 138 - 161 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes IGN] attention (apprentissage automatique) [Termes IGN] corrélation [Termes IGN] données lidar [Termes IGN] données localisées 3D [Termes IGN] figure géométrique [Termes IGN] fonction de perte [Termes IGN] graphe [Termes IGN] Perceptron multicouche [Termes IGN] scène urbaine [Termes IGN] segmentation sémantique [Termes IGN] semis de points Résumé : (auteur) Semantic segmentation of mobile laser scanning (MLS) point clouds can provide meaningful 3 D semantic information of urban facilities for various applications. However, it still remains a challenge to extract accurate 3 D semantic information from MLS point cloud data due to its irregular 3 D geometric structure in a large-scale outdoor scene. To this end, this study develops a geometry-aware attention point network (GAANet) with geometric properties of the point cloud as a reference. Specifically, the proposed method first builds a graph-like region for each input point to establish the geometric correlation toward its neighbors for robustly descripting local geometry-aware features. Thereafter, the method introduces a novel multi-head attention mechanism to efficiently learn local discriminative features on the constructed graphs and a feature combination operation to capture both local and global geometric dependencies inside fused point features for significantly facilitating the segmentation of small or incomplete 3 D objects at point-level. Finally, an adaptive loss function is appended to handle class imbalance for the overall performance improvement. The validation experiments on two challenging benchmarks demonstrate the effectiveness and powerful generation ability of the proposed method, which achieves state-of-the-art performance with mean IoU of 65.09% and 95.20% in the Toronto-3D and Oakland 3-D MLS dataset, respectively. Numéro de notice : A2023-038 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1080/13658816.2022.2111572 Date de publication en ligne : 24/08/2022 En ligne : https://doi.org/10.1080/13658816.2022.2111572 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=102309 [article]

ChineseTR: A weakly supervised toponym recognition architecture based on automatic training data generator and deep neural network / Qinjun Qiu in Transactions in GIS, vol 26 n° 3 (May 2022)  

Public [article]  inTransactions in GIS > vol 26 n° 3 (May 2022) . - pp 1256 - 1279 Titre : ChineseTR: A weakly supervised toponym recognition architecture based on automatic training data generator and deep neural network Type de document : Article/Communication Auteurs : Qinjun Qiu, Auteur ; Zhong Xie, Auteur ; Shu Wang, Auteur ; et al., Auteur Année de publication : 2022 Article en page(s) : pp 1256 - 1279 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géomatique web [Termes IGN] apprentissage profond [Termes IGN] Chine [Termes IGN] classification par réseau neuronal récurrent [Termes IGN] données d'entrainement (apprentissage automatique) [Termes IGN] données issues des réseaux sociaux [Termes IGN] échantillonnage de données [Termes IGN] OpenStreetMap [Termes IGN] reconnaissance automatique [Termes IGN] répertoire toponymique [Termes IGN] site wiki [Termes IGN] toponyme Résumé : (auteur) Toponym recognition is used to extract toponyms from natural language texts, which is a fundamental task of ubiquitous geographic information applications. Existing toponym recognition methods with state-of-the-art performance mainly leverage supervised learning (i.e., deep-learning-based approaches) with parameters learned from massive, labeled datasets that must be annotated manually. This is a great inconvenience when model training needs to fit different domain texts, especially those of social media messaging. To address this issue, this article proposes a weakly supervised Chinese toponym recognition (ChineseTR) architecture that leverages a training dataset creator that generates training datasets automatically based on word collections and associated word frequencies from various texts and an extension recognizer that employs a basic bidirectional recurrent neural network based on particular features designed for toponym recognition. The results show that the proposed ChineseTR achieves a 0.76 F1 score in a corpus with a 0.718 out-of-vocabulary rate and a 0.903 in-vocabulary rate. All comparative experiments demonstrate that ChineseTR is an effective and scalable architecture that recognizes toponyms. Numéro de notice : A2022-462 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article DOI : 10.1111/tgis.12902 Date de publication en ligne : 02/02/2022 En ligne : https://doi.org/10.1111/tgis.12902 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100796 [article]

Spatially oriented convolutional neural network for spatial relation extraction from natural language texts / Qinjun Qiu in Transactions in GIS, vol 26 n° 2 (April 2022)  

Public [article]  inTransactions in GIS > vol 26 n° 2 (April 2022) . - pp 839 - 866 Titre : Spatially oriented convolutional neural network for spatial relation extraction from natural language texts Type de document : Article/Communication Auteurs : Qinjun Qiu, Auteur ; Zhong Xie, Auteur ; Kai Ma, Auteur ; et al., Auteur Année de publication : 2022 Article en page(s) : pp 839 - 866 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géomatique web [Termes IGN] appariement sémantique [Termes IGN] apprentissage dirigé [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] données d'entrainement (apprentissage automatique) [Termes IGN] exploration de données [Termes IGN] langage naturel (informatique) [Termes IGN] proximité sémantique [Termes IGN] relation spatiale [Termes IGN] relation topologique [Termes IGN] site wiki [Termes IGN] spatial metrics [Termes IGN] système à base de connaissances Résumé : (auteur) Spatial relation extraction (e.g., topological relations, directional relations, and distance relations) from natural language descriptions is a fundamental but challenging task in several practical applications. Current state-of-the-art methods rely on rule-based metrics, either those specifically developed for extracting spatial relations or those integrated in methods that combine multiple metrics. However, these methods all rely on developed rules and do not effectively capture the characteristics of natural language spatial relations because the descriptions may be heterogeneous and vague and may be context sparse. In this article, we present a spatially oriented piecewise convolutional neural network (SP-CNN) that is specifically designed with these linguistic issues in mind. Our method extends a general piecewise convolutional neural network with a set of improvements designed to tackle the task of spatial relation extraction. We also propose an automated workflow for generating training datasets by integrating new sentences with those in a knowledge base, based on string similarity and semantic similarity, and then transforming the sentences into training data. We exploit a spatially oriented channel that uses prior human knowledge to automatically match words and understand the linguistic clues to spatial relations, finally leading to an extraction decision. We present both the qualitative and quantitative performance of the proposed methodology using a large dataset collected from Wikipedia. The experimental results demonstrate that the SP-CNN, with its supervised machine learning, can significantly outperform current state-of-the-art methods on constructed datasets. Numéro de notice : A2022-365 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article DOI : 10.1111/tgis.12887 Date de publication en ligne : 27/12/2021 En ligne : https://doi.org/10.1111/tgis.12887 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100584 [article]