A5045885893- 3D Shape Modeling and Analysis
- Additive Manufacturing and 3D Printing Technologies
- Manufacturing Process and Optimization
- Computer Graphics and Visualization Techniques
- Advanced Numerical Analysis Techniques
- Coral and Marine Ecosystems Studies
- Lung Cancer Diagnosis and Treatment
- Marine animal studies overview
- Jewish and Middle Eastern Studies
- Architecture and Computational Design
- Experimental Learning in Engineering
- COVID-19 diagnosis using AI
- Ichthyology and Marine Biology
- Remote Sensing and LiDAR Applications
- Anatomy and Medical Technology
- Maritime and Coastal Archaeology
- Computational Geometry and Mesh Generation
- Underwater Acoustics Research
- Structural Analysis and Optimization
- 3D Surveying and Cultural Heritage
- Engineering Technology and Methodologies
- Education and Technology Integration
- Domain Adaptation and Few-Shot Learning
- Mobile Learning in Education
- Archaeology and Historical Studies
Istituto di Scienza e Tecnologie dell'Informazione "Alessandro Faedo"
2018-2023
University of Cagliari
2016-2018
New York University
2018
Abstract Semantic segmentation is a widespread image analysis task; in some applications, it requires such high accuracy that still has to be done manually, taking long time. Deep learning‐based approaches can significantly reduce times, but current automated solutions may produce results below expert standards. We propose agLab, an interactive tool for the rapid labelling and of orthoimages speeds up semantic segmentation. TagLab follows human‐centered artificial intelligence approach that,...
Abstract We propose a novel method for the automatic generation of structured hexahedral meshes articulated 3D shapes. recast complex problem generating connectivity mesh general shape into simpler tubular structure derived from its curve‐skeleton. also provide volumetric subdivision schemes to nicely adapt topology local thickness tubes, while regularizing per‐element size. Our is fast, one‐click, easy reproduce, and it generates that better align branching input if compared previous...
We propose a novel algorithm for decomposing general three-dimensional geometries into small set of overlap-free height-field blocks , volumes enclosed by flat base and surface defined with respect to this base. This decomposition is useful fabrication methodologies such as 3-axis CNC milling, where single milling pass can only carve the machine tray but also benefit other settings. Computing our desired requires solving highly constrained discrete optimization problem, variants which are...
Detecting the impacts of natural and anthropogenic disturbances that cause declines in organisms or changes community composition has long been a focus ecology. However, tradeoff often exists between spatial extent over which relevant data can be collected, resolution those data. Recent advances underwater photogrammetry, as well computer vision machine learning tools employ artificial intelligence (AI), offer potential solutions with to resolve this tradeoff. Here, we coupled rigorous...
Abstract We introduce a novel geometry‐processing pipeline to guide the fabrication of complex shapes from single block material using 4‐axis CNC milling machines. This setup extends classical 3‐axis machining with an extra degree freedom rotate object around fixed axis. The first step our identifies rotation axis that maximizes overall accuracy. Then we identify two height‐field regions at axis's extremes used secure on tool. segment remaining portion mesh into set height‐fields whose...
Mesostructured patterns are a modern and efficient concept based on designing the geometry of structural material at meso-scale to achieve desired mechanical performances. In context bending-active structures, such can be used control flexibility panels forming surface without changing constituting material. These undergo formation process deformation by bending, application internal restraints. This paper describes new constructional system, FlexMaps, that has initiated adoption...
Fully automatic semantic segmentation of highly specific classes and complex shapes may not meet the accuracy standards demanded by scientists. In such cases, human-centered AI solutions, able to assist operators while preserving human control over tasks, are a good trade-off speed up image labeling maintaining high levels. TagLab is an open-source AI-assisted software for annotating large orthoimages which takes advantage different degrees automation; it speeds annotation from scratch...