A5066942213- Retinal Imaging and Analysis
- Seismic Imaging and Inversion Techniques
- Glaucoma and retinal disorders
- Offshore Engineering and Technologies
- Reservoir Engineering and Simulation Methods
- Retinal Diseases and Treatments
- Computational Physics and Python Applications
- Marine and Offshore Engineering Studies
- Stock Market Forecasting Methods
- Wave and Wind Energy Systems
- Methane Hydrates and Related Phenomena
- Time Series Analysis and Forecasting
- Oil and Gas Production Techniques
- Semantic Web and Ontologies
- Hydrological Forecasting Using AI
- Model Reduction and Neural Networks
- Underwater Acoustics Research
- Seismic Waves and Analysis
- Service-Oriented Architecture and Web Services
- Complex Systems and Time Series Analysis
- Drilling and Well Engineering
- Neural Networks and Applications
- Underwater Vehicles and Communication Systems
- Oceanographic and Atmospheric Processes
- Seismology and Earthquake Studies
Universidade de São Paulo
2010-2024
Universidade Federal de São Paulo
2022
Hospital Universitário da Universidade de São Paulo
2021-2022
Universidade Politecnica
2022
Institute of Electrical and Electronics Engineers
2018
National Day Laborer Organizing Network
2018
Purpose . To investigate the diagnostic accuracy of machine learning classifiers (MLCs) using retinal nerve fiber layer (RNFL) and optic (ON) parameters obtained with spectral domain optical coherence tomography (SD-OCT). Methods Fifty-seven patients early to moderate primary open angle glaucoma 46 healthy were recruited. All 103 underwent a complete ophthalmological examination, achromatic standard automated perimetry, imaging SD-OCT. Receiver operating characteristic (ROC) curves built for...
To evaluate the sensitivity and specificity of machine learning classifiers (MLCs) for glaucoma diagnosis using Spectral Domain OCT (SD-OCT) standard automated perimetry (SAP).Observational cross-sectional study. Sixty two patients 48 healthy individuals were included. All underwent a complete ophthalmologic examination, achromatic (SAP) retinal nerve fiber layer (RNFL) imaging with SD-OCT (Cirrus HD-OCT; Carl Zeiss Meditec Inc., Dublin, California). Receiver operating characteristic (ROC)...
The mooring systems give stability to the floating platforms against environmental conditions, stabilizing platform with lines attached seabed. are among main components that guarantee safety of staff and various operations carried out on platforms. current approaches used monitor inefficient as line tension sensors expensive install, maintain, have durability problems. This article presents development two neural network-based machine learning systems: a Multilayer Perceptron (MLP) Long...
To test the ability of machine learning classifiers (MLCs) using optical coherence tomography (OCT) and standard automated perimetry (SAP) parameters to discriminate between healthy glaucomatous individuals, compare it diagnostic combined structure-function index (CSFI), general ophthalmologists glaucoma specialists.Cross-sectional prospective study.Fifty eight eyes 58 patients with early moderate (median value mean deviation = -3.44 dB; interquartile range, -6.0 -2.4 dB) 66 individuals...
Purpose. To investigate the sensitivity and specificity of machine learning classifiers (MLC) spectral domain optical coherence tomography (SD-OCT) for diagnosis glaucoma. Methods. Sixty-two patients with early to moderate glaucomatous visual field damage 48 healthy individuals were included. All subjects underwent a complete ophthalmologic examination, achromatic standard automated perimetry, RNFL imaging SD-OCT (Cirrus HD-OCT; Carl Zeiss Meditec, Inc., Dublin, California, USA). Receiver...
Abstract We develop and implement a neural operator (NOp) to predict the evolution of waves on surface water. The NOp uses graph network (GNN) connect randomly sampled points water exchange information between them make prediction. Our main contribution is adding physical knowledge implementation, which allows model be more general able used in domains different geometries with no retraining. implementation also takes advantage fact that governing equations are independent rotation...
The current design process of mooring systems for Floating Production, Storage, and Offloading units (FPSOs) depends on the availability platform's mathematical model accuracy dynamic simulations. These simulations then provide FPSO's time series motion which is evaluated according to constraints. This can be time-consuming present inaccurate results due model's limitations overall complexity vessel's dynamics. We propose a Neural Simulator, called NeuroSim, set data-based surrogate models...
Researchers typically resort to numerical methods understand and predict ocean dynamics, a key task in mastering environmental phenomena. Such may not be suitable scenarios where the topographic map is complex, knowledge about underlying processes incomplete, or application time critical. On other hand, if dynamics are observed, they can exploited by recent machine learning methods. In this paper we describe data-driven method variables such as current velocity sea surface height region of...
Abstract The current design process of mooring systems for FPSOs is highly dependent on the availability platform’s mathematical model and accuracy dynamic simulations, through which resulting time series motion evaluated according to constraints. This can be time-consuming present inaccurate results due model’s limitations overall complexity vessel’s dynamics. We propose a Neural Simulator, set data-based surrogate models with environmental data as input, each specialized in prediction...
Sea-level rise is a well-known consequence of climate change. Several studies have estimated the social and economic impact increase in extreme flooding. An efficient way to mitigate its consequences development flood alert prediction system, based on high-resolution numerical models robust sensing networks. However, current use various simplifying assumptions that compromise accuracy ensure solvability within reasonable timeframe, hindering more regular cost-effective forecasts for...
Abstract This paper presents a Neural Operator (NOp) for estimating wave height from wind data. The NOp is machine learning model to approximate functions one function space another. It trained map data height. Each represented as graph, with vertices corresponding physical locations and edges carrying geographical distribution information. uses convolutional kernel, allowing nodes exchange information based on relative positions. After several convolutions, the latent values are decoded...
Abstract Robust monitoring is vital in offshore oil and gas exploration, especially for Floating Production, Storage, Offloading (FPSO) platforms pivotal subsea hydrocarbon extraction. Motivated by the glaring absence of failure detection systems mooring lines on FPSOs, this work tackles broader issue inadequate warning sector despite their role safety. Existing approaches often overlook inter-series correlations multivariate scenarios. The authors introduce Graph Neural Networks (GNNs),...
The simulation of acousitc wave propagation is the kernel for important industrial applications like Full-Waveform Inversion (FWI) and Reverse-Time Migration (RTM). solves partial differential equations (PDEs) based on finite differences method, which can be significantly accelerated with support GPUs. One main challenges accelerating this stencil computations GPUs to reduce overhead memory accesses, tiling an optimization accelerate kernels. However, deciding tile sizes these not a...
Abstract Mooring line breakage of a moored offshore Floating Production Storage and Offloading (FPSO) platform results in an increased load on the remaining lines, worsening their degradation rates. Furthermore, break causes FPSO to change its motions move away from desired location. In addition, differ depending draft environmental conditions, making it difficult detect mooring failure. this paper, we propose system based machine learning called Neural Motion Estimator (NeMo) composed (i)...
Glaucoma is an optical neuropathy, whose progression results in visual field impairments and blindness. In this paper artificial data generator called GLOR presented, which based on a Monte Carlo method designed for the training of machine learning classifiers glaucoma diagnosis. The generated population characterized by functional structural eyes. study, these parameters are provided high definition coherence tomography (HD-OCT) standard automated perimetry (SAP) instruments. A Naive-Bayes...
This paper proposes a solution based on Multi-Layer Perceptron (MLP) to predict the offset of center gravity an offshore platform. It also performs comparative study with three optimization algorithms – Random Search, Simulated Annealing, and Bayesian Optimization (BO) find best MLP architecture. Although BO obtained architecture in shortest time, ablation studies developed this hyperparameters process showed that result is sensitive them deserves attention Neural Architecture Search process.
Multivariate Time Series Classification (MTSC) is a complex problem that has seen great advances in recent years from the application of state-of-the-art machine learning techniques. However, there still need for thorough evaluation effect signal noise classification performance MTSC To this end, paper, we evaluate three current and effective classifiers – DDTW, ROCKET InceptionTime propose their use real-world problem: detection mooring line failure offshore platforms. We show all them...