A5031358325- Robotics and Sensor-Based Localization
- Underwater Vehicles and Communication Systems
- Underwater Acoustics Research
- Advanced Image and Video Retrieval Techniques
- Advanced Vision and Imaging
- Legal and Labor Studies
- Water Quality Monitoring Technologies
- Dental Implant Techniques and Outcomes
- Health and Medical Education
- 3D Surveying and Cultural Heritage
- Dental Radiography and Imaging
- Image Enhancement Techniques
- Color Science and Applications
- Inertial Sensor and Navigation
- Advanced Image Fusion Techniques
- Italian Social Issues and Migration
- Human Rights and Immigration
- Oral microbiology and periodontitis research
- Human Pose and Action Recognition
- Remote Sensing and LiDAR Applications
- Oral and gingival health research
- Occupational Health and Safety in Workplaces
- Oropharyngeal Anatomy and Pathologies
- Environmental Monitoring and Data Management
- Augmented Reality Applications
Ifremer
2021-2024
Université Paris-Saclay
2017-2021
Université de Montpellier
2019
Centre National de la Recherche Scientifique
2019
Laboratoire d'Informatique, de Robotique et de Microélectronique de Montpellier
2018-2019
Universidad de Sevilla
2004-2010
Many applications of Visual SLAM, such as augmented reality, virtual robotics or autonomous driving, require versatile, robust and precise solutions, most often with real-time capability. In this work, we describe OV <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> a fully online algorithm, handling both monocular stereo camera setups, various map scales frame-rates ranging from few Hertz up to several hundreds. It combines numerous...
In the context of underwater robotics, visual degradation induced by medium properties make difficult exclusive use cameras for localization purpose. Hence, many methods are based on expensive navigation sensors associated with acoustic positioning. On other hand, pure have shown great potential in but challenging conditions, such as presence turbidity and dynamism, remain complex to tackle. this paper, we propose a new odometry method designed be robust these perturbations. The proposed...
Underwater images are altered by the physical characteristics of medium through which light rays pass before reaching optical sensor. Scattering and wavelength-dependent absorption significantly modify captured colors depending on distance observed elements to image plane. In this paper, we aim recover an scene as if water had no effect propagation. We introduce SUCRe, a novel method that exploits scene's 3D structure for underwater color restoration. By following points in multiple tracking...
This paper addresses the lack of “push-button” software for optical marine imaging, which currently limits use photogrammetric approaches by a wider community. It presents and reviews an open source software, Matisse, creating textured 3D models complex underwater scenes from video or still images. developed non-experts, enables routine efficient processing images into that facilitate exploitation analysis imagery. When vehicle navigation data are available, Matisse allows seamless...
Visual localization plays an important role in the positioning and navigation of robotics systems within previously visited environments. When visits occur over long periods time, changes environment related to seasons or day-night cycles present a major challenge. Under water, sources variability are due other factors such as water conditions growth marine organisms. Yet, it remains obstacle much less studied one, partly lack data. This paper presents new deep-sea dataset benchmark...
Imaging is increasingly used to capture information on the marine environment thanks improvements in imaging equipment, devices for carrying cameras and data storage recent years. In that context, biologists, geologists, computer specialists end-users must gather discuss methods procedures optimising quality quantity of collected from images. The 4 th Marine Workshop was organised 3-6 October 2022 Brest (France) a hybrid mode. More than hundred participants were welcomed person about 80...
We present a new dataset, dedicated to the development of simultaneous localization and mapping methods for underwater vehicles navigating close seabed. The data sequences composing this dataset are recorded in three different environments: harbor at depth few meters, first archeological site 270 second 380 meters. acquisition is performed using remotely operated equipped with monocular monochromatic camera, low-cost inertial measurement unit, pressure sensor, computing all embedded single...
Hyperspectral imaging has been increasingly used for underwater survey applications over the past years. As many hyperspectral cameras work as push-broom scanners, their use is usually limited to creation of photo-mosaics based on a flat surface approximation and by interpolating camera pose from dead-reckoning navigation. Yet, because drift in navigation mostly wrong assumption, quality obtained often too low support adequate analysis. In this paper we present an initial method creating...
The estimation of disparity maps from stereo pairs has many applications in robotics and autonomous driving. Stereo matching first been solved using model-based approaches, with real-time considerations for some, but today's most recent works rely on deep convolutional neural networks mainly focus accuracy at the expense computing time. In this paper, we present a new method getting best both worlds: data-based methods speed fast ones. proposed approach fuses prior to estimate refined...
This paper presents a new underwater dataset acquired from visual-inertial-pressure acquisition system and meant to be used benchmark visual odometry, SLAM multi-sensors solutions. The is publicly available contains ground-truth trajectories for evaluation.
de implantes transicionales en sobredentaduras mandibulares adultos mayores.Av
Underwater images are altered by the physical characteristics of medium through which light rays pass before reaching optical sensor. Scattering and wavelength-dependent absorption significantly modify captured colors depending on distance observed elements to image plane. In this paper, we aim recover an scene as if water had no effect propagation. We introduce SUCRe, a novel method that exploits scene's 3D structure for underwater color restoration. By following points in multiple tracking...
Research in 3D mapping is crucial for smart city applications, yet the cost of acquiring data often hinders progress. Visual localization, particularly monocular camera position estimation, offers a solution by determining camera's pose solely through visual cues. However, this task challenging due to limited from single camera. To tackle these challenges, we organized AISG–SLA Localization Challenge (VLC) at IJCAI 2023 explore how AI can accurately extract 2D images space. The challenge...
Research in 3D mapping is crucial for smart city applications, yet the cost of acquiring data often hinders progress. Visual localization, particularly monocular camera position estimation, offers a solution by determining camera's pose solely through visual cues. However, this task challenging due to limited from single camera. To tackle these challenges, we organized AISG-SLA Localization Challenge (VLC) at IJCAI 2023 explore how AI can accurately extract 2D images space. The challenge...
This paper presents a combination of real-time and offline 3D reconstruction methods for remotely operated vehicles (ROVs) equipped with cameras used in underwater inspection survey tasks. The component is based on stereo visual simultaneous localization mapping algorithm truncated signed distance field representation producing coarse online. uses structure-from-motion techniques to create dense point cloud the scene which then meshed textured produce high-quality mesh. highlights...
Many applications of Visual SLAM, such as augmented reality, virtual robotics or autonomous driving, require versatile, robust and precise solutions, most often with real-time capability. In this work, we describe OV$^{2}$SLAM, a fully online algorithm, handling both monocular stereo camera setups, various map scales frame-rates ranging from few Hertz up to several hundreds. It combines numerous recent contributions in visual localization within an efficient multi-threaded architecture....
Introducción.El objetivo del presente estudio era mostrar los resultados de tratamiento con implantes dentales insertados inmediatamente después la extracción.Métodos.22 pacientes pérdida dental unitaria, parcial o total fueron tratados 82 Microdent ® superficie chorreado arena y grabada ácidos.Todos extracción correspondiente.Los cargados un periodo cicatrización 3 meses en mandíbula 6 el maxilar superior.Resultados.Los hallazgos clínicos indican una supervivencia éxito 97,6%. 2 se...
Hyperspectral (HS) imaging produces an image of object across a large range the visible spectrum, and not just primary colors (R, G, B) conventional cameras. It can provide valuable information for detection, analysis materials processes in environmental science deep-sea, especially study benthic environments pollution monitoring. In this paper, we address problem camera calibration towards 3D hyperspectral mapping where GPS is available, platform navigational sensors are accurate enough to...
Introducción.El estudio muestra la evaluación de los pacientes edéntulos totales tratados con sobredentaduras mandibulares implantes dentales mediante diferentes tipos carga funcional.Métodos.33 fueron 127 en mandíbula para su rehabilitación prostodóncica sobredentaduras.Todos insertados cirugía y cargados forma convencional, precoz o inmediata.Resultados.Después un seguimiento clínico medio 67,5 meses (rango: 36-108 meses), resultados indican éxito del 96,9% porque 4 fracasaron.El 100% una...
This paper is a technical report about our submission for the ECCV 2018 3DRMS Workshop Challenge on Semantic 3D Reconstruction \cite{Tylecek2018rms}. In this paper, we address semantic reconstruction autonomous navigation using co-learning of depth map and segmentation. The core pipeline deep multi-task neural network which tightly refines also produces accurate segmentation maps. Its inputs are an image raw produced from pair images by standard stereo vision. resulting point clouds then...