Centre dedocumentation
scientifique

Connecting images through sources: Exploring low-data, heterogeneous instance retrieval / Dimitri Gominski in Remote sensing, vol 13 n° 16 (August-2 2021)  

Public [article]  inRemote sensing > vol 13 n° 16 (August-2 2021) . - n° 3080 Titre : Connecting images through sources: Exploring low-data, heterogeneous instance retrieval Type de document : Article/Communication Auteurs : Dimitri Gominski , Auteur ; Valérie Gouet-Brunet , Auteur ; Liming Chen, Auteur Année de publication : 2021 Projets : Alegoria / Gouet-Brunet, Valérie Article en page(s) : n° 3080 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] description multiniveau [Termes IGN] patrimoine culturel [Termes IGN] recherche d'image basée sur le contenu [Termes IGN] test de performance Résumé : (auteur) Along with a new volume of images containing valuable information about our past, the digitization of historical territorial imagery has brought the challenge of understanding and interconnecting collections with unique or rare representation characteristics, and sparse metadata. Content-based image retrieval offers a promising solution in this context, by building links in the data without relying on human supervision. However, while the latest propositions in deep learning have shown impressive results in applications linked to feature learning, they often rely on the hypothesis that there exists a training dataset matching the use case. Increasing generalization and robustness to variations remains an open challenge, poorly understood in the context of real-world applications. Introducing the alegoria benchmark, containing multi-date vertical and oblique aerial digitized photography mixed with more modern street-level pictures, we formulate the problem of low-data, heterogeneous image retrieval, and propose associated evaluation setups and measures. We propose a review of ideas and methods to tackle this problem, extensively compare state-of-the-art descriptors and propose a new multi-descriptor diffusion method to exploit their comparative strengths. Our experiments highlight the benefits of combining descriptors and the compromise between absolute and cross-domain performance. Numéro de notice : A2021-610 Affiliation des auteurs : UGE-LaSTIG+Ext (2020- ) Thématique : IMAGERIE/INFORMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.3390/rs13163080 date de publication en ligne : 05/08/2021 En ligne : https://doi.org/10.3390/rs13163080 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98357 [article]

Detection of pictorial map objects with convolutional neural networks / Raimund Schnürer in Cartographic journal (the), vol 58 n° 1 (August 2021)  

Public [article]  inCartographic journal (the) > vol 58 n° 1 (August 2021) . - pp 50 - 68 Titre : Detection of pictorial map objects with convolutional neural networks Type de document : Article/Communication Auteurs : Raimund Schnürer, Auteur ; René Sieber, Auteur ; Jost Schmid-Lanter, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : pp 50 - 68 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] base de données d'images [Termes IGN] bibliothèque numérique [Termes IGN] carte ancienne [Termes IGN] carte numérique [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] détection d'objet [Termes IGN] données issues des réseaux sociaux [Termes IGN] objet cartographique [Termes IGN] pictogramme Résumé : (auteur) In this work, realistically drawn objects are identified on digital maps by convolutional neural networks. For the first two experiments, 6200 images were retrieved from Pinterest. While alternating image input options, two binary classifiers based on Xception and InceptionResNetV2 were trained to separate maps and pictorial maps. Results showed that the accuracy is 95–97% to distinguish maps from other images, whereas maps with pictorial objects are correctly classified at rates of 87–92%. For a third experiment, bounding boxes of 3200 sailing ships were annotated in historic maps from different digital libraries. Faster R-CNN and RetinaNet were compared to determine the box coordinates, while adjusting anchor scales and examining configurations for small objects. A resulting average precision of 32% was obtained for Faster R-CNN and of 36% for RetinaNet. Research outcomes are relevant for trawling map images on the Internet and for enhancing the advanced search of digital map catalogues. Numéro de notice : A2021-651 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/00087041.2020.1738112 date de publication en ligne : 11/09/2020 En ligne : https://doi.org/10.1080/00087041.2020.1738112 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98381 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
030-2021011	SL	Revue	Centre de documentation	Revues en salle	Disponible

Deep learning in denoising of micro-computed tomography images of rock samples / Mikhail Sidorenko in Computers & geosciences, vol 151 (June 2021)  

Public [article]  inComputers & geosciences > vol 151 (June 2021) . - n° 104716 Titre : Deep learning in denoising of micro-computed tomography images of rock samples Type de document : Article/Communication Auteurs : Mikhail Sidorenko, Auteur ; Denis Orlov, Auteur ; Mohammad Ebadi, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : n° 104716 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] accentuation d'image [Termes IGN] apprentissage profond [Termes IGN] classification dirigée [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] filtrage du bruit [Termes IGN] filtre passe-bande [Termes IGN] roche [Termes IGN] tomographie Résumé : (auteur) Nowadays, the advantages of Digital Rock Physics (DRP) are well known and widely applied in comprehensive core analysis. It is also known that the quality of the 3D pore scale model drastically influences the results of rock properties simulation, which makes the preprocessing stage of DRP very important. In this work, we consider the application of Deep Convolutional Neural Networks (CNNs) for the preprocessing of CT images, specifically for denoising, in two setups - conventional fully-supervised learning and the self-supervised learning, when the only available data is the noisy images. To train CNNs in a supervised setup, we use images processed by a combination of bilateral and bandpass filters. We trained CNNs of the same architecture with different loss functions to find out how the choice of a loss function influences the model's performance. Some of the obtained CNNs yielded the highest quality in terms of full-reference and no-reference metrics and significant histogram effect (bimodal intensity distribution). Images denoised with these models were qualitatively and quantitatively better than the reference “ground truth” images used for training. We use the Deep Image Prior algorithm to train denoising models in a self-supervised setup. The obtained models are much better than ones obtained in fully-supervised setup, but are too slow, as they are optimization-based rather than feed-forward. Such an algorithm can be used in the dataset generation for feed-forward meta-models. These results could help to develop an AI-based instrument to build high-quality 3D segmented models of rocks for DRP applications. Numéro de notice : A2021-389 Affiliation des auteurs : non IGN Thématique : IMAGERIE/INFORMATIQUE Nature : Article DOI : 10.1016/j.cageo.2021.104716 date de publication en ligne : 02/03/2021 En ligne : https://doi.org/10.1016/j.cageo.2021.104716 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97672 [article]

Direct analysis in real-time (DART) time-of-flight mass spectrometry (TOFMS) of wood reveals distinct chemical signatures of two species of Afzelia / Peter Kitin in Annals of Forest Science [en ligne], vol 78 n° 2 (June 2021)  

Public [article]  inAnnals of Forest Science [en ligne] > vol 78 n° 2 (June 2021) . - Article 31 Titre : Direct analysis in real-time (DART) time-of-flight mass spectrometry (TOFMS) of wood reveals distinct chemical signatures of two species of Afzelia Type de document : Article/Communication Auteurs : Peter Kitin, Auteur ; Edgard Espinoza, Auteur ; Hans Beeckman, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : Article 31 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] abattage (sylviculture) [Termes IGN] Afzelia (genre) [Termes IGN] analyse discriminante [Termes IGN] apprentissage non-dirigé [Termes IGN] bois [Termes IGN] espèce végétale [Termes IGN] forêt tropicale [Termes IGN] identification de plantes [Termes IGN] signature spectrale [Termes IGN] spectrométrie [Termes IGN] taxinomie [Termes IGN] temps réel Résumé : (Auteur) Distinct chemical fingerprints of the wood of Afzelia pachyloba and A. bipindensis demonstrated an effective method for identifying these two commercially important species. Direct analysis in real-time (DART) time-of-flight mass spectrometry (TOFMS) allowed high-throughput examination of chemotypes with vast potential in taxonomic, ecological, and forensic research of wood. Context : Afzelia is a genus of valuable tropical timber trees. Accurate identification of wood is required for the prevention of illicit timber trade as well as for certification purposes in the forest and wood products industry. For many years, particular interest has been focused on attempts to distinguish the wood of A. bipindensis Harms from A. pachyloba Harms due to substantial differences in the commercial values of these two species. Aims : We investigated if wood chemical signatures and microscopy could identify the wood of A. bipindensis and A. pachyloba. Methods : We used two approaches, namely metabolome profiling by direct analysis in real-time (DART) time-of-flight mass spectrometry (TOFMS) and wood microstructure by light microscopy and SEM. In all, we analyzed samples from 89 trees of A. bipindensis, and A. pachyloba. Results : The two species could not be separated by the IAWA standard microscopic wood features. SEM analysis showed considerable variation in the morphology of vestured pits; however, this variation was not species-specific. In contrast, DART-TOFMS followed by unsupervised statistics (Discriminant Analysis of Principal Components) showed distinct metabolome signatures of the two species. Conclusion : DART-TOFMS provides a rapid method for wood identification that can be easily applied to small heartwood samples. Time- and cost-effective classification of wood chemotypes by DART-TOFMS can have potential applications in various research questions in forestry, wood science, tree-ecophysiology, and forensics. Numéro de notice : A2021-327 Affiliation des auteurs : non IGN Thématique : FORET/IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1007/s13595-020-01024-1 date de publication en ligne : 31/03/2021 En ligne : https://doi.org/10.1007/s13595-020-01024-1 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97488 [article]

Multiple convolutional features in Siamese networks for object tracking / Zhenxi Li in Machine Vision and Applications, vol 32 n° 3 (May 2021)  

Public [article]  inMachine Vision and Applications > vol 32 n° 3 (May 2021) . - n° 59 Titre : Multiple convolutional features in Siamese networks for object tracking Type de document : Article/Communication Auteurs : Zhenxi Li, Auteur ; Guillaume-Alexandre Bilodeau, Auteur ; Wassim Bouachir, Auteur Année de publication : 2021 Article en page(s) : n° 59 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] approche hiérarchique [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] extraction de traits caractéristiques [Termes IGN] poursuite de cible [Termes IGN] reconnaissance d'objets [Termes IGN] réseau neuronal siamois Résumé : (auteur) Siamese trackers demonstrated high performance in object tracking due to their balance between accuracy and speed. Unlike classification-based CNNs, deep similarity networks are specifically designed to address the image similarity problem and thus are inherently more appropriate for the tracking task. However, Siamese trackers mainly use the last convolutional layers for similarity analysis and target search, which restricts their performance. In this paper, we argue that using a single convolutional layer as feature representation is not an optimal choice in a deep similarity framework. We present a Multiple Features-Siamese Tracker (MFST), a novel tracking algorithm exploiting several hierarchical feature maps for robust tracking. Since convolutional layers provide several abstraction levels in characterizing an object, fusing hierarchical features allows to obtain a richer and more efficient representation of the target. Moreover, we handle the target appearance variations by calibrating the deep features extracted from two different CNN models. Based on this advanced feature representation, our method achieves high tracking accuracy, while outperforming the standard siamese tracker on object tracking benchmarks. The source code and trained models are available at https://github.com/zhenxili96/MFST. Numéro de notice : A2021-470 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1007/s00138-021-01185-7 date de publication en ligne : 11/03/2021 En ligne : https://doi.org/10.1007/s00138-021-01185-7 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97903 [article]

A stacked dense denoising–segmentation network for undersampled tomograms and knowledge transfer using synthetic tomograms / Dimitrios Bellos in Machine Vision and Applications, vol 32 n° 3 (May 2021)  

Permalink

Parsing of urban facades from 3D point clouds based on a novel multi-view domain / Wei Wang in Photogrammetric Engineering & Remote Sensing, PERS, vol 87 n° 4 (April 2021)  

Permalink

Digital surface model refinement based on projected images / Jiali Wang in Photogrammetric Engineering & Remote Sensing, PERS, vol 87 n° 3 (March 2021)  

Permalink

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

Permalink

Evaluating surface mesh reconstruction of open scenes / Yanis Marchand (2021)  

Permalink

Wavelet theory / Somayeh Mohammady (2021)  

Permalink

Crater detection and registration of planetary images through marked point processes, multiscale decomposition, and region-based analysis / David Solarna in IEEE Transactions on geoscience and remote sensing, vol 58 n° 9 (September 2020)  

Permalink

CSVM architectures for pixel-wise object detection in high-resolution remote sensing images / Youyou Li in IEEE Transactions on geoscience and remote sensing, vol 58 n° 9 (September 2020)  

Permalink

Heliport detection using artificial neural networks / Emre Baseski in Photogrammetric Engineering & Remote Sensing, PERS, vol 86 n° 9 (September 2020)  

Permalink

A novel deep network and aggregation model for saliency detection / Ye Liang in The Visual Computer, vol 36 n° 9 (September 2020)  

Permalink