Centre de documentation
scientifique

A robust nonrigid point set registration framework based on global and intrinsic topological constraints / Guiqiang Yang in The Visual Computer, vol 38 n° 2 (February 2022)  

Public [article]  inThe Visual Computer > vol 38 n° 2 (February 2022) . - pp 603 - 623 Titre : A robust nonrigid point set registration framework based on global and intrinsic topological constraints Type de document : Article/Communication Auteurs : Guiqiang Yang, Auteur ; Rui Li, Auteur ; Yujun Liu, Auteur ; Ji Wang, Auteur Année de publication : 2022 Article en page(s) : pp 603 - 623 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] algorithme espérance-maximisation [Termes IGN] contrainte géométrique [Termes IGN] contrainte topologique [Termes IGN] descripteur local [Termes IGN] méthode fondée sur le noyau [Termes IGN] méthode robuste [Termes IGN] processus gaussien [Termes IGN] semis de points [Termes IGN] superposition de données Résumé : (auteur) The problem of registering nonrigid point sets, with the aim of estimating the correspondences and learning the transformation between two given sets of points, often arises in computer vision tasks. This paper proposes a novel method for performing nonrigid point set registration on data with various types of degradation, in which the registration problem is formulated as a Gaussian mixture model (GMM)-based density estimation problem. Specifically, two complementary constraints are jointly considered for optimization in a GMM probabilistic framework. The first is a thin-plate spline-based regularization constraint that maintains global spatial motion consistency, and the second is a spectral graph-based regularization constraint that preserves the intrinsic structure of a point set. Moreover, the correspondences and the transformation are alternately optimized using the expectation maximization algorithm to obtain a closed-form solution. We first utilize local descriptors to construct the initial correspondences and then estimate the underlying transformation under the GMM-based framework. Experimental results on contour images and real images show the effectiveness and robustness of the proposed method. Numéro de notice : A2022-146 Affiliation des auteurs : non IGN Thématique : IMAGERIE/INFORMATIQUE Nature : Article DOI : 10.1007/s00371-020-02037-7 Date de publication en ligne : 21/02/2022 En ligne : https://doi.org/10.1007/s00371-020-02037-7 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100040 [article]

A convolutional neural network with mapping layers for hyperspectral image classification / Rui Li in IEEE Transactions on geoscience and remote sensing, vol 58 n° 5 (May 2020)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 58 n° 5 (May 2020) . - pp 3136 - 3147 Titre : A convolutional neural network with mapping layers for hyperspectral image classification Type de document : Article/Communication Auteurs : Rui Li, Auteur ; Zhibin Pan, Auteur ; Yang Wang, Auteur ; et al., Auteur Année de publication : 2020 Article en page(s) : pp 3136 - 3147 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] algèbre linéaire [Termes IGN] analyse discriminante [Termes IGN] analyse en composantes principales [Termes IGN] analyse multidimensionnelle [Termes IGN] apprentissage profond [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] couche thématique [Termes IGN] dispersion [Termes IGN] extraction de traits caractéristiques [Termes IGN] image hyperspectrale [Termes IGN] réduction Résumé : (auteur) In this article, we propose a convolutional neural network with mapping layers (MCNN) for hyperspectral image (HSI) classification. The proposed mapping layers map the input patch into a low-dimensional subspace by multilinear algebra. We use our mapping layers to reduce the spectral and spatial redundancies and maintain most energy of the input. The feature extracted by our mapping layers can also reduce the number of following convolutional layers for feature extraction. Our MCNN architecture avoids the declining accuracy with increasing layers phenomenon of deep learning models for HSI classification and also saves the training time for its effective mapping layers. Furthermore, we impose the 3-D convolutional kernel on the convolutional layer to extract the spectral–spatial features for HSI. We tested our MCNN on three data sets of Indian Pines, University of Pavia, and Salinas, and we achieved the classification accuracy of 98.3%, 99.5%, and 99.3%, respectively. Experimental results demonstrate that the proposed MCNN can significantly improve classification accuracy and save much time consumption. Numéro de notice : A2020-234 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2019.2948865 Date de publication en ligne : 12/11/2019 En ligne : https://doi.org/10.1109/TGRS.2019.2948865 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94980 [article]