Descripteur
Termes IGN > mathématiques > statistique mathématique > analyse de données > classification > classification par réseau neuronal > classification par réseau neuronal convolutif
classification par réseau neuronal convolutifVoir aussi |
Documents disponibles dans cette catégorie (336)
Ajouter le résultat dans votre panier
Visionner les documents numériques
Affiner la recherche Interroger des sources externes
Etendre la recherche sur niveau(x) vers le bas
Titre : Scene understanding and gesture recognition for human-machine interaction Type de document : Thèse/HDR Auteurs : Naina Dhingra, Auteur Editeur : Zurich : Eidgenossische Technische Hochschule ETH - Ecole Polytechnique Fédérale de Zurich EPFZ Année de publication : 2022 Note générale : Bibliographie
A dissertation submitted to attain the degree of Doctor of Sciences of ETH ZurichLangues : Français (fre) Descripteur : [Vedettes matières IGN] Intelligence artificielle
[Termes IGN] apprentissage profond
[Termes IGN] attention (apprentissage automatique)
[Termes IGN] classification orientée objet
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] classification par séparateurs à vaste marge
[Termes IGN] compréhension de l'image
[Termes IGN] image RVB
[Termes IGN] interaction homme-machine
[Termes IGN] oculométrie
[Termes IGN] reconnaissance automatique
[Termes IGN] reconnaissance de formes
[Termes IGN] reconnaissance de gestes
[Termes IGN] réseau neuronal récurrent
[Termes IGN] scène
[Termes IGN] vision par ordinateurRésumé : (auteur) Scene understanding and gesture recognition are useful for a myriad of applications such as human-robotic interaction, assisting blind and visually impaired people, advanced driver assistance systems, and autonomous driving. To work autonomously in real-world environments, automatic systems need to deliver non-verbal information to enhance the verbal communication in particular for blind people. We are exploring the holistic approach for providing the scene as well as gesture related information. We propose that incorporating attention mechanisms in neural networks which behave similarly to attention in the human brain, and conducting an integrated study using neural networks in real-time can yield significant improvements in the scene and gesture understanding, thereby enhancing the user experience. In this thesis, we investigate the understanding of visual scenes and gestures. We explore these two areas, in particular, by proposing novel architectures, training methods, user studies, and thorough evaluations. We show that, for deep learning approaches, attention or self attention mechanisms improve and push the boundaries of network performance for different tasks in consideration. We suggest that the various kinds of gestures can complement and supplement each other’s information to better understand non-verbal conversation; hence integrated gestures comprehension is useful. First, we focus on visual scene understanding using scene graph generation. We propose, BGT-Net, a new network that uses an object detection model with 1) bidirectional gated recurrent units for object-object communication and 2) transformer encoders including self attention to classify the objects and their relationships. We address the problem of bias caused by the long tailed distribution in the dataset. This enables the network to perform even for the unseen objects or relationships in the dataset. Second, we propose to learn hand gesture recognition from RGB and RGB-D videos using attention learning. We present a novel architecture based on residual connections and an attention mechanism. Our approach successfully detects hand gestures when evaluated on three open-source datasets. Third, we explore pointing gesture recognition and localization using open-source software, i.e. OpenPtrack which uses a deep learning based iii network to track multi-persons in the scene. We use a Kinect sensor as an input device and conduct a user study with 26 users to evaluate the system using two setup types. Fourth, we propose a technique to perform eye gaze tracking using OpenFace which is based on a deep learning model and RGB webcam. We use support vector machine regression to estimate the position of eye gaze on the screen. In a study, we evaluate the system with 28 users and show that this system can perform similarly to commercially expensive eye trackers. Finally, we focus on 3D head pose estimation using two models: 1)headPosr includes residual connections for the base network followed by a transformer encoder. It outperforms existing models but has a drawback of being computationally expensive; 2) lwPosr uses depthwise separable convolutions and transformer encoders. It is a two stream network in fine-grained fashion to estimate the three angles of the head pose. We demonstrate that this method is able to predict head poses better than state-of-the-art lightweight networks. Note de contenu : 1- Introduction
2- Background
3- State of the art
4- Scene graph generation
5- 3D hand gesture recognition
6- Pointing gesture recognition
7- Eye-gaze tracking
8- Head pose estimation
9- Lightweight head pose estimation
10- SummaryNuméro de notice : 24039 Affiliation des auteurs : non IGN Thématique : IMAGERIE/INFORMATIQUE Nature : Thèse étrangère Note de thèse : PhD Thesis : Sciences : ETH Zurich :2022 DOI : sans En ligne : https://www.research-collection.ethz.ch/handle/20.500.11850/559347 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101876 Self-attention and generative adversarial networks for algae monitoring / Nhut Hai Huynh in European journal of remote sensing, vol 55 n° 1 (2022)
[article]
Titre : Self-attention and generative adversarial networks for algae monitoring Type de document : Article/Communication Auteurs : Nhut Hai Huynh, Auteur ; Gordon Boër, Auteur ; Hauke Schramm, Auteur Année de publication : 2022 Article en page(s) : pp 10 - 22 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] algue
[Termes IGN] analyse en composantes principales
[Termes IGN] apprentissage profond
[Termes IGN] attention (apprentissage automatique)
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] image hyperspectrale
[Termes IGN] plancton
[Termes IGN] réseau antagoniste génératif
[Termes IGN] réseau neuronal artificielRésumé : (auteur) Water is important for the natural environment and human health. Monitoring algae concentrations yield information on the water quality. Compared with in situ measurements of water quality parameters, which are often complex and expensive, remote sensing techniques, using hyperspectral data analysis, are fast and cost-effective. The objectives of this study are (1) to estimate the algae concentrations from hyperspectral data using deep learning techniques, (2) to investigate the applicability of attention mechanisms in the analysis of hyperspectral data, and (3) to augment the training data using generative adversarial networks (GANs). The results show that the accuracy of deep learning techniques is 7.6% higher than that of simpler artificial neural networks. Compared to noise injection and principal component analysis-based data augmentation, the use of a GAN-based data augmentation method significantly improves the accuracy of algae concentration estimates (>5%). In addition, models with added attention mechanisms yield an on average 3.13% higher accuracy than those without attention techniques. This result demonstrates the improvement of spectral features of artificial hyperspectral data based on the self-attention approach, revealing the potential of attention techniques in hyperspectral remote sensing. Numéro de notice : A2022-097 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1080/22797254.2021.2010605 Date de publication en ligne : 02/01/2022 En ligne : https://doi.org/10.1080/22797254.2021.2010605 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99547
in European journal of remote sensing > vol 55 n° 1 (2022) . - pp 10 - 22[article]Semantic segmentation of high-resolution remote sensing images based on a class feature attention mechanism fused with Deeplabv3+ / Zhimin Wang in Computers & geosciences, vol 158 (January 2022)
[article]
Titre : Semantic segmentation of high-resolution remote sensing images based on a class feature attention mechanism fused with Deeplabv3+ Type de document : Article/Communication Auteurs : Zhimin Wang, Auteur ; Jiasheng Wang, Auteur ; Kun Yang, Auteur ; et al., Auteur Année de publication : 2022 Article en page(s) : n° 104969 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] apprentissage profond
[Termes IGN] attention (apprentissage automatique)
[Termes IGN] classe sémantique
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] image à haute résolution
[Termes IGN] image Gaofen
[Termes IGN] raisonnement sémantique
[Termes IGN] segmentation sémantiqueRésumé : (auteur) Aiming at solving the problems of inaccurate segmentation of edge targets, inconsistent segmentation of different types of targets, and slow prediction efficiency on semantic segmentation of high-resolution remote sensing images by classical semantic segmentation network, this study proposed a class feature attention mechanism fused with an improved Deeplabv3+ network called CFAMNet for semantic segmentation of common features in remote sensing images. First, the correlation between classes is enhanced using the class feature attention module to extract and process different categories of semantic information better. Second, the multi-parallel atrous spatial pyramid pooling structure is used to enhance the correlation between spaces, to extract the context information of different scales of an image better. Finally, the encoder-decoder structure is used to refine the segmentation results. The segmentation effect of the proposed network is verified by experiments on the public data set GaoFen image dataset (GID). The experimental results show that the CFAMNet can achieve the mean intersection over union (MIOU) and overall accuracy (OA) of 77.22% and 85.01%, respectively, on the GID, thus surpassing the current mainstream semantic segmentation networks. Numéro de notice : A2022-030 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1016/j.cageo.2021.104969 Date de publication en ligne : 26/10/2021 En ligne : https://doi.org/10.1016/j.cageo.2021.104969 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99269
in Computers & geosciences > vol 158 (January 2022) . - n° 104969[article]Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany / Omar Seleem in Geomatics, Natural Hazards and Risk, vol 13 (2022)
[article]
Titre : Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany Type de document : Article/Communication Auteurs : Omar Seleem, Auteur ; Georgy Ayzel, Auteur Année de publication : 2022 Article en page(s) : pp 1640 - 1662 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Analyse spatiale
[Termes IGN] Berlin
[Termes IGN] cartographie des risques
[Termes IGN] classification par forêts d'arbres décisionnels
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] classification par séparateurs à vaste marge
[Termes IGN] inondation
[Termes IGN] pouvoir de résolution géométrique
[Termes IGN] vulnérabilitéRésumé : (auteur) Identifying urban pluvial flood-prone areas is necessary but the application of two-dimensional hydrodynamic models is limited to small areas. Data-driven models have been showing their ability to map flood susceptibility but their application in urban pluvial flooding is still rare. A flood inventory (4333 flooded locations) and 11 factors which potentially indicate an increased hazard for pluvial flooding were used to implement convolutional neural network (CNN), artificial neural network (ANN), random forest (RF) and support vector machine (SVM) to: (1) Map flood susceptibility in Berlin at 30, 10, 5, and 2 m spatial resolutions. (2) Evaluate the trained models' transferability in space. (3) Estimate the most useful factors for flood susceptibility mapping. The models' performance was validated using the Kappa, and the area under the receiver operating characteristic curve (AUC). The results indicated that all models perform very well (minimum AUC = 0.87 for the testing dataset). The RF models outperformed all other models at all spatial resolutions and the RF model at 2 m spatial resolution was superior for the present flood inventory and predictor variables. The majority of the models had a moderate performance for predictions outside the training area based on Kappa evaluation (minimum AUC = 0.8). Aspect and altitude were the most influencing factors on the image-based and point-based models respectively. Data-driven models can be a reliable tool for urban pluvial flood susceptibility mapping wherever a reliable flood inventory is available. Numéro de notice : A2022-457 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE/INFORMATIQUE Nature : Article DOI : 10.1080/19475705.2022.2097131 Date de publication en ligne : 12/07/2022 En ligne : https://doi.org/10.1080/19475705.2022.2097131 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101257
in Geomatics, Natural Hazards and Risk > vol 13 (2022) . - pp 1640 - 1662[article]Efficient occluded road extraction from high-resolution remote sensing imagery / Dejun Feng in Remote sensing, vol 13 n° 24 (December-2 2021)
[article]
Titre : Efficient occluded road extraction from high-resolution remote sensing imagery Type de document : Article/Communication Auteurs : Dejun Feng, Auteur ; Xingyu Shen, Auteur ; Yakun Xie, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : n° 4974 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 de partie cachée
[Termes IGN] extraction de traits caractéristiques
[Termes IGN] extraction du réseau routier
[Termes IGN] image à haute résolution
[Termes IGN] reconstruction de routeRésumé : (auteur) Road extraction is important for road network renewal, intelligent transportation systems and smart cities. This paper proposes an effective method to improve road extraction accuracy and reconstruct the broken road lines caused by ground occlusion. Firstly, an attention mechanism-based convolution neural network is established to enhance feature extraction capability. By highlighting key areas and restraining interference features, the road extraction accuracy is improved. Secondly, for the common broken road problem in the extraction results, a heuristic method based on connected domain analysis is proposed to reconstruct the road. An experiment is carried out on a benchmark dataset to prove the effectiveness of this method, and the result is compared with that of several famous deep learning models including FCN8s, SegNet, U-Net and D-Linknet. The comparison shows that this model increases the IOU value and the F1 score by 3.35–12.8% and 2.41–9.8%, respectively. Additionally, the result proves the proposed method is effective at extracting roads from occluded areas. Numéro de notice : A2021-889 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.3390/rs13244974 Date de publication en ligne : 07/12/2021 En ligne : https://doi.org/10.3390/rs13244974 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99243
in Remote sensing > vol 13 n° 24 (December-2 2021) . - n° 4974[article]Automatic extraction of indoor spatial information from floor plan image: A patch-based deep learning methodology application on large-scale complex buildings / Hyunjung Kim in ISPRS International journal of geo-information, vol 10 n° 12 (December 2021)PermalinkBuilding detection with convolutional networks trained with transfer learning / Simon Šanca in Geodetski vestnik, vol 65 n° 4 (December 2021 - February 2022)PermalinkDiResNet: Direction-aware residual network for road extraction in VHR remote sensing images / Lei Ding in IEEE Transactions on geoscience and remote sensing, vol 59 n° 12 (December 2021)PermalinkLithological mapping based on fully convolutional network and multi-source geological data / Ziye Wang in Remote sensing, vol 13 n° 23 (December-1 2021)PermalinkMSegnet, a practical network for building detection from high spatial resolution images / Bo Yu in Photogrammetric Engineering & Remote Sensing, PERS, vol 87 n° 12 (December 2021)PermalinkVGI3D: an interactive and low-cost solution for 3D building modelling from street-level VGI images / Chaoquan Zhang in Journal of Geovisualization and Spatial Analysis, vol 5 n° 2 (December 2021)PermalinkA CNN-based approach for the estimation of canopy heights and wood volume from GEDI waveforms / Ibrahim Fayad in Remote sensing of environment, vol 265 (November 2021)PermalinkFully automated pose estimation of historical images in the context of 4D geographic information systems utilizing machine learning methods / Ferdinand Maiwald in ISPRS International journal of geo-information, vol 10 n° 11 (November 2021)PermalinkMulti-objective CNN-based algorithm for SAR despeckling / Sergio Vitale in IEEE Transactions on geoscience and remote sensing, vol 59 n° 11 (November 2021)PermalinkPose estimation and 3D reconstruction of vehicles from stereo-images using a subcategory-aware shape prior / Maximilian Alexander Coenen in ISPRS Journal of photogrammetry and remote sensing, Vol 181 (November 2021)Permalink