Listing 1 - 10 of 45 | << page >> |
Sort by
|
Choose an application
Alzheimer’s disease (AD) is one of the most common neurodegenerative diseases in the world and the most common cause of dementia. In recent times, accurate and early detection of AD plays a vital role in patient care and further treatment. Lately, studies on AD diagnosis has attached a great significanceto artificial-based diagnostic algorithms. During this master thesis we explore how deep learning models can handle neuroimages in order to identify andpredict the evolution of the disease. Different from the traditional machine learning algorithms, deep learning does not require manually extracted features but instead utilizes 3D image processing models to learn features for the diagnosis and the prognosis of AD. The contribution of this work relies on a more rigorous preprocessing phase involving skull-stripping and intensity normalization of the medical images. The hippocampus, a brain area critical for learning and memory, is especially affected at early stages of Alzheimer’s disease. In some parts of this work, It will be used as a region of interest for our algorithms that will consist in convolutional neural networks, the typical image classifier models, and vision transformers, a novel deep learning architecture.
Choose an application
Deep learning has been revolutionizing many fields in computer vision, and facial informatics is one of the major fields. Novel approaches and performance breakthroughs are often reported on existing benchmarks. As the performances on existing benchmarks are close to saturation, larger and more challenging databases are being made and considered as new benchmarks, further pushing the advancement of the technologies. Considering face recognition, for example, the VGG-Face2 and Dual-Agent GAN report nearly perfect and better-than-human performances on the IARPA Janus Benchmark A (IJB-A) benchmark. More challenging benchmarks, e.g., the IARPA Janus Benchmark A (IJB-C), QMUL-SurvFace and MegaFace, are accepted as new standards for evaluating the performance of a new approach. Such an evolution is also seen in other branches of face informatics. In this Special Issue, we have selected the papers that report the latest progresses made in the following topics: 1. Face liveness detection 2. Emotion classification 3. Facial age estimation 4. Facial landmark detection We are hoping that this Special Issue will be beneficial to all fields of facial informatics.
History of engineering & technology --- deep learning --- RGB --- depth --- facial landmarking --- merging networks --- 3D geometry data --- 2D attribute maps --- fused CNN feature --- coarse-to-fine --- convolutional neural network (CNN) --- deep metric learning --- multi-task learning --- image classification --- age estimation --- generative adversarial network --- emotion classification --- facial key point detection --- facial images processing --- convolutional neural networks --- face liveness detection --- convolutional neural network --- thermal image --- external knowledge
Choose an application
Les services écosystémiques récréatifs sont généralement mal évalués ou ignorés. Une meilleure évaluation de ces services devrait permettre aux gestionnaires de milieux naturels de comprendre plus finement la dynamique des visiteurs dans ces milieux. Cela aiderait à mieux lutter contre la dégradation des milieux naturels, mais aussi à mettre certains écosystèmes plus en valeur. Dans ce mémoire, un prototype d’une nouvelle technique d’évaluation de la fréquentation touristique est élaboré. Cette technique utilise les camera traps pour collecter des données sur le nombre de visiteurs et sur leur comportement. Le travail d’analyse de ces images, souvent conséquent, est confié à un algorithme de deep learning nommé Mask R-CNN. D’une part, les résultats montrent que la position des caméras traps doit être homogénéisée pour faciliter le fonctionnement de l’algorithme. D’autre part, le modèle détecte 89,0 ± 4,2 % du temps les personnes et les classifie correctement 97,1 ± 1,0 % du temps sur des photos floues. Néanmoins, les autres classes (chiens, vélos, véhicules, sacs à dos) ne sont pas encore correctement détectées. Le modèle doit être partiellement entraîné sur des données en extérieur et un prétraitement des images par segmentation devrait être utilisé. Une étude de cas réalisée sur le projet AGRETA montre que, même si la technique doit être améliorée, il est déjà possible d’obtenir certaines données précises sur le long terme. De plus, pouvoir consulter les images lors d’événements suspects dans les résultats est une source d’informations inattendue, mais riche. Recreational ecosystem services are generally poorly valued or ignored. However, a better evaluation of these services should allow managers to better understand the dynamics of visitors in these environments. This would help to better prevent the degradation of natural environments but also to highlight certain ecosystems. In this work, a prototype of a new technique for visitor monitoring is being developed. This technique uses camera traps to collect counting and behavioural data of visitors. The analysing task of these images, often substantial, is carried out by a deep learning algorithm called Mask R-CNN. The results show, on the one hand, that the position of the camera traps must be homogenized to facilitate the operation of the algorithm. On the other hand, the model detects people 89.0 ± 4.2% of the time and classifies them correctly 97.1 ± 1.0% of the time on blurred photos. However, the other classes (dogs, bicycles, vehicles, backpacks) are not correctly detected yet. The model must be partially trained on outdoor data, and image pre-processing by segmentation should be used. A case study of the AGRETA project shows that although the technique needs to be improved, it is already possible to obtain some accurate data over the long term. In addition, the ability to view images during suspicious events in the results is an unexpected but rich source of information.
services écosystémiques culturels, réseaux neuronaux convolutionnels, CNN, Mask R-CNN, piège caméra, fréquentation touristique, nombre de visiteurs, écotourisme, Ardenne, parcs nationaux --- cultural ecosystem services, convolutional neural network, CNN, Mask R-CNN, camera trap, visitor monitoring, visitor number, ecotourism, Ardenne, national parks --- Sciences du vivant > Sciences de l'environnement & écologie --- Ingénierie, informatique & technologie > Sciences informatiques --- Sciences sociales & comportementales, psychologie > Géographie humaine & démographie
Choose an application
Deep learning has been revolutionizing many fields in computer vision, and facial informatics is one of the major fields. Novel approaches and performance breakthroughs are often reported on existing benchmarks. As the performances on existing benchmarks are close to saturation, larger and more challenging databases are being made and considered as new benchmarks, further pushing the advancement of the technologies. Considering face recognition, for example, the VGG-Face2 and Dual-Agent GAN report nearly perfect and better-than-human performances on the IARPA Janus Benchmark A (IJB-A) benchmark. More challenging benchmarks, e.g., the IARPA Janus Benchmark A (IJB-C), QMUL-SurvFace and MegaFace, are accepted as new standards for evaluating the performance of a new approach. Such an evolution is also seen in other branches of face informatics. In this Special Issue, we have selected the papers that report the latest progresses made in the following topics: 1. Face liveness detection 2. Emotion classification 3. Facial age estimation 4. Facial landmark detection We are hoping that this Special Issue will be beneficial to all fields of facial informatics.
History of engineering & technology --- deep learning --- RGB --- depth --- facial landmarking --- merging networks --- 3D geometry data --- 2D attribute maps --- fused CNN feature --- coarse-to-fine --- convolutional neural network (CNN) --- deep metric learning --- multi-task learning --- image classification --- age estimation --- generative adversarial network --- emotion classification --- facial key point detection --- facial images processing --- convolutional neural networks --- face liveness detection --- convolutional neural network --- thermal image --- external knowledge
Choose an application
Deep learning has been revolutionizing many fields in computer vision, and facial informatics is one of the major fields. Novel approaches and performance breakthroughs are often reported on existing benchmarks. As the performances on existing benchmarks are close to saturation, larger and more challenging databases are being made and considered as new benchmarks, further pushing the advancement of the technologies. Considering face recognition, for example, the VGG-Face2 and Dual-Agent GAN report nearly perfect and better-than-human performances on the IARPA Janus Benchmark A (IJB-A) benchmark. More challenging benchmarks, e.g., the IARPA Janus Benchmark A (IJB-C), QMUL-SurvFace and MegaFace, are accepted as new standards for evaluating the performance of a new approach. Such an evolution is also seen in other branches of face informatics. In this Special Issue, we have selected the papers that report the latest progresses made in the following topics: 1. Face liveness detection 2. Emotion classification 3. Facial age estimation 4. Facial landmark detection We are hoping that this Special Issue will be beneficial to all fields of facial informatics.
deep learning --- RGB --- depth --- facial landmarking --- merging networks --- 3D geometry data --- 2D attribute maps --- fused CNN feature --- coarse-to-fine --- convolutional neural network (CNN) --- deep metric learning --- multi-task learning --- image classification --- age estimation --- generative adversarial network --- emotion classification --- facial key point detection --- facial images processing --- convolutional neural networks --- face liveness detection --- convolutional neural network --- thermal image --- external knowledge
Choose an application
The recent revolution in deep learning has enabled considerable development in the fields of object and activity detection. Visual object detection tries to find objects of target classes with precise localisation in an image and assign each object instance a corresponding class label. At the same time, activity recognition aims to determine the actions or activities of an agent or group of agents based on sensor or video observation data. It is a very important and challenging problem to detect, identify, track, and understand the behaviour of objects through images and videos taken by various cameras. Together, objects and their activity recognition in imaging data captured by remote sensing platforms is a highly dynamic and challenging research topic. During the last decade, there has been significant growth in the number of publications in the field of object and activity recognition. In particular, many researchers have proposed application domains to identify objects and their specific behaviours from air and spaceborne imagery. This Special Issue includes papers that explore novel and challenging topics for object and activity detection in remote sensing images and videos acquired by diverse platforms.
multi-camera system --- space alignment --- UAV-assisted calibration --- cross-view matching --- spatiotemporal feature map --- view-invariant description --- air-to-ground synchronization --- tidal flat water --- YOLOv3 --- similarity algorithm for water extraction --- arbitrary-oriented object detection in satellite optical imagery --- adaptive dynamic refined single-stage transformer detector --- feature pyramid transformer --- dynamic feature refinement --- synthetic aperture radar (SAR) --- ship detection --- convolutional neural network (CNN) --- deep learning (DL) --- feature pyramid network (FPN) --- quad feature pyramid network (Quad-FPN) --- crowd estimation --- 3D simulation --- unmanned aerial vehicle --- synthetic crowd data --- invasive species --- thermal imaging --- habitat identification --- deep learning --- drone --- multiview semantic vegetation index --- urban forestry --- green view index (GVI) --- semantic segmentation --- urban vegetation --- RGB vegetation index --- n/a
Choose an application
Tribology has been and continues to be one of the most relevant fields, being present in almost all aspects of our lives. The understanding of tribology provides us with solutions for future technical challenges. At the root of all advances made so far are multitudes of precise experiments and an increasing number of advanced computer simulations across different scales and multiple physical disciplines. Based upon this sound and data-rich foundation, advanced data handling, analysis and learning methods can be developed and employed to expand existing knowledge. Therefore, modern machine learning (ML) or artificial intelligence (AI) methods provide opportunities to explore the complex processes in tribological systems and to classify or quantify their behavior in an efficient or even real-time way. Thus, their potential also goes beyond purely academic aspects into actual industrial applications. To help pave the way, this article collection aimed to present the latest research on ML or AI approaches for solving tribology-related issues generating true added value beyond just buzzwords. In this sense, this Special Issue can support researchers in identifying initial selections and best practice solutions for ML in tribology.
artificial intelligence --- machine learning --- artificial neural networks --- tribology --- condition monitoring --- semi-supervised learning --- random forest classifier --- self-lubricating journal bearings --- reduced order modelling --- dynamic friction --- rubber seal applications --- tensor decomposition --- laser surface texturing --- texturing during moulding --- digital twin --- PINN --- reynolds equation --- triboinformatics --- databases --- data mining --- meta-modeling --- monitoring --- analysis --- prediction --- optimization --- fault data generation --- Convolutional Neural Network (CNN) --- Generative Adversarial Network (GAN) --- bearing fault diagnosis --- unbalanced datasets --- tribo-testing --- tribo-informatics --- natural language processing --- tribAIn --- BERT --- amorphous carbon coatings --- UHWMPE --- total knee replacement --- Gaussian processes --- rolling bearing dynamics --- cage instability --- regression --- neural networks --- random forest --- gradient boosting --- evolutionary algorithms --- rolling bearings --- remaining useful life --- feature engineering --- structure-borne sound --- n/a
Choose an application
The recent revolution in deep learning has enabled considerable development in the fields of object and activity detection. Visual object detection tries to find objects of target classes with precise localisation in an image and assign each object instance a corresponding class label. At the same time, activity recognition aims to determine the actions or activities of an agent or group of agents based on sensor or video observation data. It is a very important and challenging problem to detect, identify, track, and understand the behaviour of objects through images and videos taken by various cameras. Together, objects and their activity recognition in imaging data captured by remote sensing platforms is a highly dynamic and challenging research topic. During the last decade, there has been significant growth in the number of publications in the field of object and activity recognition. In particular, many researchers have proposed application domains to identify objects and their specific behaviours from air and spaceborne imagery. This Special Issue includes papers that explore novel and challenging topics for object and activity detection in remote sensing images and videos acquired by diverse platforms.
Technology: general issues --- History of engineering & technology --- multi-camera system --- space alignment --- UAV-assisted calibration --- cross-view matching --- spatiotemporal feature map --- view-invariant description --- air-to-ground synchronization --- tidal flat water --- YOLOv3 --- similarity algorithm for water extraction --- arbitrary-oriented object detection in satellite optical imagery --- adaptive dynamic refined single-stage transformer detector --- feature pyramid transformer --- dynamic feature refinement --- synthetic aperture radar (SAR) --- ship detection --- convolutional neural network (CNN) --- deep learning (DL) --- feature pyramid network (FPN) --- quad feature pyramid network (Quad-FPN) --- crowd estimation --- 3D simulation --- unmanned aerial vehicle --- synthetic crowd data --- invasive species --- thermal imaging --- habitat identification --- deep learning --- drone --- multiview semantic vegetation index --- urban forestry --- green view index (GVI) --- semantic segmentation --- urban vegetation --- RGB vegetation index --- n/a
Choose an application
Unmanned aerial vehicles (UAVs) are new platforms that have been increasingly used in the last few years for forestry applications that benefit from the added value of flexibility, low cost, reliability, autonomy, and capability of timely provision of high-resolution data. The main adopted image-based technologies are RGB, multispectral, and thermal infrared. LiDAR sensors are becoming commonly used to improve the estimation of relevant plant traits. In comparison with other permanent ecosystems, forests are particularly affected by climatic changes due to the longevity of the trees, and the primary objective is the conservation and protection of forests. Nevertheless, forestry and agriculture involve the cultivation of renewable raw materials, with the difference that forestry is less tied to economic aspects and this is reflected by the delay in using new monitoring technologies. The main forestry applications are aimed toward inventory of resources, map diseases, species classification, fire monitoring, and spatial gap estimation. This Special Issue focuses on new technologies (UAV and sensors) and innovative data elaboration methodologies (object recognition and machine vision) for applications in forestry.
Research & information: general --- Biology, life sciences --- Forestry & related industries --- unmanned aerial vehicles --- seedling detection --- forest regeneration --- reforestation --- establishment survey --- machine learning --- multispectral classification --- UAV photogrammetry --- forest modeling --- ancient trees measurement --- tree age prediction --- Mauritia flexuosa --- semantic segmentation --- end-to-end learning --- convolutional neural network --- forest inventory --- Unmanned Aerial Systems (UAS) --- structure from motion (SfM) --- Unmanned Aerial Vehicles (UAV) --- Photogrammetry --- Thematic Mapping --- Accuracy Assessment --- Reference Data --- Forest Sampling --- Remote Sensing --- Robinia pseudoacacia L. --- reproduction --- spreading --- short rotation coppice --- unmanned aerial system (UAS) --- object-based image analysis (OBIA) --- convolutional neural network (CNN) --- juniper woodlands --- ecohydrology --- remote sensing --- unmanned aerial systems --- central Oregon --- rangelands --- seedling stand inventorying --- photogrammetric point clouds --- hyperspectral imagery --- leaf-off --- leaf-on --- UAV --- multispectral image --- forest fire --- burn severity --- classification --- precision agriculture --- biomass evaluation --- image processing --- Castanea sativa --- unmanned aerial vehicles (UAV) --- precision forestry --- forestry applications --- RGB imagery
Choose an application
The concept of remote sensing as a way of capturing information from an object without making contact with it has, until recently, been exclusively focused on the use of Earth observation satellites.The emergence of unmanned aerial vehicles (UAV) with Global Navigation Satellite System (GNSS) controlled navigation and sensor-carrying capabilities has increased the number of publications related to new remote sensing from much closer distances. Previous knowledge about the behavior of the Earth's surface under the incidence different wavelengths of energy has been successfully applied to a large amount of data recorded from UAVs, thereby increasing the special and temporal resolution of the products obtained.More specifically, the ability of UAVs to be positioned in the air at pre-programmed coordinate points; to track flight paths; and in any case, to record the coordinates of the sensor position at the time of the shot and at the pitch, yaw, and roll angles have opened an interesting field of applications for low-altitude aerial photogrammetry, known as UAV photogrammetry. In addition, photogrammetric data processing has been improved thanks to the combination of new algorithms, e.g., structure from motion (SfM), which solves the collinearity equations without the need for any control point, producing a cloud of points referenced to an arbitrary coordinate system and a full camera calibration, and the multi-view stereopsis (MVS) algorithm, which applies an expanding procedure of sparse set of matched keypoints in order to obtain a dense point cloud. The set of technical advances described above allows for geometric modeling of terrain surfaces with high accuracy, minimizing the need for topographic campaigns for georeferencing of such products.This Special Issue aims to compile some applications realized thanks to the synergies established between new remote sensing from close distances and UAV photogrammetry.
Technology: general issues --- unmanned aerial vehicle --- urban LULC --- GEOBIA --- multiscale classification --- unmanned aircraft system (UAS) --- deep learning --- super-resolution (SR) --- convolutional neural network (CNN) --- generative adversarial network (GAN) --- structure-from-motion --- photogrammetry --- remote sensing --- UAV --- 3D-model --- surveying --- vertical wall --- snow --- remotely piloted aircraft systems --- structure from motion --- lidar --- forests --- orthophotography --- construction planning --- sustainable construction --- urbanism --- BIM --- building maintenance --- unmanned aerial vehicle (UAV) --- structure-from-motion (SfM) --- ground control points (GCP) --- accuracy assessment --- point clouds --- corridor mapping --- UAV photogrammetry --- terrain modeling --- vegetation removal --- unmanned aerial vehicles --- power lines --- image-based reconstruction --- 3D reconstruction --- unmanned aerial systems --- time series --- accuracy --- reproducibility --- orthomosaic --- validation --- drone --- GNSS RTK --- precision --- elevation --- multispectral imaging --- vegetation indices --- nutritional analysis --- correlation --- optimal harvest time --- UAV images --- monoscopic mapping --- stereoscopic plotting --- image overlap --- optimal image selection --- n/a
Listing 1 - 10 of 45 | << page >> |
Sort by
|