A5021255741- Advanced Vision and Imaging
- Video Analysis and Summarization
- 3D Surveying and Cultural Heritage
- Industrial Vision Systems and Defect Detection
- Video Surveillance and Tracking Methods
- Robotics and Sensor-Based Localization
- Image Retrieval and Classification Techniques
- Image and Object Detection Techniques
- Anomaly Detection Techniques and Applications
- Manufacturing Process and Optimization
- Infrastructure Maintenance and Monitoring
- Non-Destructive Testing Techniques
- Cell Image Analysis Techniques
- Cultural Heritage Materials Analysis
- Human Pose and Action Recognition
- Geographic Information Systems Studies
- Soil and Land Suitability Analysis
- Remote Sensing and LiDAR Applications
- Digital Holography and Microscopy
- Color Science and Applications
- Welding Techniques and Residual Stresses
- Sports Analytics and Performance
- Land Use and Ecosystem Services
- Maritime Navigation and Safety
- Advanced Optical Sensing Technologies
National Research Council
2015-2024
Tecnologie Avanzate (Italy)
2023-2024
Institute of Intelligent Industrial Systems and Technologies for Advanced Manufacturing
2020-2023
Institute of Intelligent Systems for Automation
2005-2022
ETH Zurich
2016
University of South Australia
2011
National Academies of Sciences, Engineering, and Medicine
2009
National Research Council Canada
2007
United States Nuclear Regulatory Commission
2007
The problem of ground truth generation is fundamental for many approaches computer vision and image processing. In order to test algorithms object segmentation, tracking, interactions, it necessary have sequences in which the determined an objective way. context visual surveillance where people moves scene occluding each other, could be very complex hard work generating position all moving objects maintain this information period they remain scene. paper we propose a semi-automatic system...
In recent years sport video research has gained a steady interest among the scientific community. The large amount of data available from broadcast transmissions and dedicated camera setups, need extracting meaningful information data, pose significant challenges. Hence, computer vision machine learning are essential for enabling automated or semi-automated processing big in sports. Although sports diverse enough to present unique challenges on their own, most them share identify active...
Abstract In‐flow phase‐contrast tomography provides a 3D refractive index of label‐free cells in cytometry systems. Its major limitation, as with any quantitative phase imaging approach, is the lack specificity compared to fluorescence microscopy, thus restraining its huge potentialities single‐cell analysis and diagnostics. Remarkable results introducing are obtained through artificial intelligence (AI), but only for adherent cells. However, accessing ground truth obtaining accurate...
In this paper, we present a new generation of omnidirectional automated guided vehicles (omniagv) used for transporting materials within manufacturing factory with the ability to navigate autonomously and intelligently by interacting environment, including people other entities. This robot has be integrated into operating environment without significant changes current facilities or heavy redefinitions logistics processes already running. For purpose, different vision-based systems advanced...
An osmotic imbalance between the two water phases of multiple water-in-oil-in-water (W1/O/W2) emulsions results in either emulsion swelling or shrinking due to migration across oil layer. Controlled mass transport is not only importance for stability but also allows transient thickening controlled release encapsulated substances, such as nutriments simply salt. Our prior work has shown that follows sequential stages. In first stage, oil-phase structure changed a way rapid, osmotically driven...
In this paper, we investigate on the feasibility of multiple camera system for automatic offside detection. We propose six fixed cameras, properly placed two sides soccer field (three each side) to reduce perspective and occlusion errors. The images acquired by synchronized cameras are processed detect players' position ball in real-time; a view analysis is carried out evaluate event, considering all players field, determining who passed ball, if active condition occurred. whole has been...
Monitoring and maintaining the health state of existing bridges is a time-consuming critical task. To reduce time effort required for first screening to prioritize risks, deep-learning-based object detectors can be used. In detail, automatic defect damage recognition on elements performed using single-stage detectors, such as YOLOv5. this end, database typical defects was gathered labeled by domain experts YOLOv5 trained, tested, validated. Results showed good effectiveness accuracy proposed...
In this paper we tackle the problem of indoor robot localization by using a vision-based approach. Specifically, propose visual odometer able to give back relative pose an omnidirectional automatic guided vehicle (AGV) that moves inside industrial environment. A monocular downward-looking camera having optical axis nearly perpendicular ground floor, is used for collecting floor images. After preliminary analysis images aimed at detecting robust point features (keypoints) takes place,...
In response to prevailing classification inconsistency between land cover maps, developed by different organizations in times at scales, an object-based National Land Representation System (NLRS) for Bangladesh has been developed. The process, which began 2013 and was completed 2016, brought together several national involved extensive process of consultation, data collection, translation, analysis existing cover/use systems. focused on the interpretation three legends from historic maps....
Information on land cover is vital to numerous United Nations (UN) missions, including achieving the Sustainable Development Goals (SDGs). Because data are developed by a variety of organizations for range objectives, they based different classification schemes and have discrepancies. In addition, sustainability hampered limited access information documentation. Accordingly, international standards improve interoperability between datasets. However, use development datasets various factors...
Land cover (LC) plays a crucial role in the monitoring and planning of environment among various domains, such as climate change, management natural resources, sustainable development. However, inconsistent LC legends hamper their usability, particularly technologies, like remote sensing, rapidly improve data quality quantity for extracting valuable information. Ontologies play pivotal improving standardization harmonization different taxonomies, enabling both reduction inconsistencies way...
In the last decade, several research efforts have been undertaken in soccer video analysis. This increasing interest is motivated by possible applications over a wide spectrum of topics: indexing, summarization, enhancement, team and players statistics, tactic analysis, referee's support, etc. Soccer analysis requires different challenging tasks: ball to be localized each frame, tracked time and, above all, their interactions detected analyzed. The latter task fundamental, especially for...
In this work we consider the problem of soccer team discrimination. The approach propose starts from monocular images acquired by a still camera. first step is player detection, performed means background subtraction. An algorithm based on pixels energy content has been implemented in order to detect moving objects. use information, combined with temporal sliding window procedure, allows be substantially independent motion hypothesis. Colour histograms RGB space are extracted each player,...
We propose a method for solving one of the significant open issues in computer vision: material recognition. A time-of-flight range camera has been employed to analyze characteristics different materials. Starting from information returned by depth sensor, features interest have extracted using transforms such as Fourier, discrete cosine, Hilbert, chirp-z, and Karhunen–Loève. Such used build training validation set useful feed classifier (J48) able accomplish recognition step. The...
Tennis player silhouette extraction is a preliminary step fundamental for any behavior analysis processing. Automatic systems the evaluation of tactics, in terms position court, postures during game and types strokes, are highly desired coaches training purposes. These require accurate segmentation players order to apply posture high level semantic analysis. Background subtraction algorithms have been largely used sportive context when fixed cameras used. In this paper an innovative...
In this paper we present a multi-people-tracking algorithm which is able to detect and track humans in complex situations with varying light conditions, high frame rate, real time processing. We propose stochastic approach for foreground people tracking based on the evaluation of maximum posteriori probability (MAP). The evaluates geometrical information blob overlapping does not require feature extraction single object. Experimental tests have been carried out soccer image sequence some...
Segmentation is a significant preliminary step for many image-based object recognition activities. Microscopy images often present segmentation problems, namely low contrast (the objects are translucent) and occlusions. Fortunately, translucency provides some possibility of solving the occlusion problem; edge-based methods can be used to tackle (translucency) problem, but edges noisy edge tracking must used. In occluded regions very faint noise conflicting confuse even tracking: an contour...
Land cover (LC) is an essential variable for environmental monitoring in many application domains. The detection of changes LC can support the understanding dynamics. However, legends present a high degree inconsistencies that significantly reduce their usability. This study investigates effectiveness ISO standard 19144-2, better known as Cover Meta-Language (LCML), to improve standardization and harmonization different taxonomies maps. LCML vocabulary syntactic rules facilitate integration...
Computer vision is steadily gaining importance in many research fields, as its applications expand from traditional fields situation analysis and scene understanding video surveillance to other scenarios. The sportive context can represent a perfect test-bed for machine algorithms because of the large availability visual data brought by wide spread cameras on relatively high number courts. In this paper we introduce tennis ball detection tracking method that exploits domain knowledge...
In the last decade, soccer video analysis has received a lot of attention from scientific community. This increasing interest is motivated by possible applications over wide spectrum topics: indexing, summarization, enhancement, team and players statistics, tactics analysis, referee support, etc. The application computer vision methodologies in context requires many problems to be faced: ball have detected images any light weather condition, they localized field, tracked time finally their...