A5051251449- Seismic Imaging and Inversion Techniques
- Remote Sensing in Agriculture
- Remote Sensing and LiDAR Applications
- Smart Agriculture and AI
- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- Drilling and Well Engineering
- Remote-Sensing Image Classification
- Medical Image Segmentation Techniques
- Seismic Waves and Analysis
- Climate variability and models
- Growth and nutrition in plants
- Meteorological Phenomena and Simulations
- Retinal Imaging and Analysis
- Machine Learning and Data Classification
- Seed Germination and Physiology
- Face recognition and analysis
- Reservoir Engineering and Simulation Methods
- Speech and Audio Processing
- Advanced Neural Network Applications
- Cloud Computing and Resource Management
- Banana Cultivation and Research
- Automated Road and Building Extraction
- Data Management and Algorithms
- Seismology and Earthquake Studies
University of Wisconsin–Madison
2024
Fundação Getulio Vargas
2023-2024
Technical University of Munich
2022-2024
Wageningen University & Research
2023
IBM Research - Brazil
2017-2022
IBM (United States)
2021
Pontifical Catholic University of Rio de Janeiro
2008-2015
Universidade de São Paulo
2015
Public Health Dayton & Montgomery County
2013
Much of the progress in development highly adaptable and reusable artificial intelligence (AI) models is expected to have a profound impact on Earth science remote sensing. Foundation are pre-trained large unlabeled datasets through self-supervision, then fine-tuned for various downstream tasks with small labeled datasets. There an increasing interest within scientific community investigate how effectively build generalist AI that exploit multi-sensor data observation applications. This...
Having dense and regularly sampled data is becoming increasingly important in seismic processing. However, due to physical or financial constraints, sets can be often undersampled. Occasionally, these may also present bad dead traces the geoscientist must deal with. Many works have tackled this problem using prestack classified three main categories: wave-equation, domain transform, prediction-error-filter methods. In letter, we assess performance of a conditional generative adversarial...
Automatic document layout analysis is a crucial step in cognitive computing and processes that extract information out of images, such as specific-domain knowledge database creation, graphs images understanding, extraction structured data from tables, others. Even with the progress observed this field last years, challenges are still open range accurately detecting content boxes to classifying them into semantically meaningful classes. With popularization mobile devices cloud-based services,...
Abstract. Optical imagery is often affected by the presence of clouds. Aiming to reduce their effects, different reconstruction techniques have been proposed in last years. A common alternative extract data from active sensors, like Synthetic Aperture Radar (SAR), because they are almost independent on atmospheric conditions and solar illumination. On other hand, SAR images more complex interpret than optical requiring particular handling. Recently, Conditional Generative Adversarial...
This paper proposes a new distributed architecture for supervised classification of large volumes earth observation data on cloud computing environment. The supports execution, network communication, and fault tolerance in transparent way to the user. is composed three abstraction layers, which support definition implementation applications by researchers from different scientific investigation fields. also discussed. A software prototype (available online), runs machine learning routines...
Noisy traces, gaps in coverage, or irregular/inadequate trace spacing are common problems both land and marine surveys, possibly hindering the geological interpretation of an area interest. This problem has been typically addressed literature using prestack data; however, data not always available. As alternative, poststack interpolations may aid by increasing spatial density a seismic section can also be used to reconstruct entire sections interpolating neighboring reducing field costs. In...
Abstract. Since many years ago, the scientific community is concerned about how to increase accuracy of different classification methods, and major achievements have been made so far. Besides this issue, increasing amount data that being generated every day by remote sensors raises more challenges be overcome. In work, a tool within scope InterIMAGE Cloud Platform (ICP), which an open-source, distributed framework for automatic image interpretation, presented. The tool, named ICP: Data...
The population growth and consequent global rise in food demand require increasingly efficient agricultural solutions, what is commonly called digital agriculture. Among promising initiatives, the use of remotely sensed data combined with machine learning algorithms enables handling faster operations lower associated cost. One most important activities agriculture crop identification, which fundamental for managing inventory a farm by producers governmental authorities, has been addressed...
The application of deep learning algorithms to Earth observation (EO) in recent years has enabled substantial progress fields that rely on remotely sensed data. However, given the data scale EO, creating large datasets with pixel-level annotations by experts is expensive and highly time-consuming. In this context, priors are seen as an attractive way alleviate burden manual labeling when training methods for EO. For some applications, those readily available. Motivated great success...
Seismic exploration is a complex process that depends on different sources of information. An essential one seismic imaging, and much its interpretation performance relies high-quality processing, which currently still very dependent prone-to-error human mediation. Automation such processing steps necessary to reduce the amount time treat data—usually months—and improve outcome overall quality by reducing inherent subjectivity in process. One most critical noise suppression, ground roll...
The use of priors to avoid manual labeling for training machine learning methods has received much attention in the last few years. One critical subthemes this regard is Learning from Label Proportions (LLP), where only information about class proportions available models. While various LLP settings verse literature, most approaches focus on bag-level label errors, often leading suboptimal solutions. This paper proposes a new model that jointly uses prototypical contrastive and cluster...
The definition of reliable velocity functions is paramount for obtaining high-quality poststack seismic data. Velocity are commonly created with the interpreter interactively selecting high-energy peaks in spectra and verifying if derived match traveltime trajectories corresponding common midpoint (CMP) gathers. Modern software further allows to apply resulting moveout corrections verify desired overall flatness reflection events achieved. This very detailed process takes a significant...
The characterization of the subsurface is paramount for exploration and production life cycle and, more specifically, identification potential hydrocarbon accumulations. In recent years, oil gas industry has increased its interest in applying machine learning to accelerate seismic interpretation process, which regarded as a time-consuming human-centered task. Although been successfully used many applications ranging from stratigraphic segmentation salt dome detection, usual bottleneck need...
Abstract The stochastic synthesis of extreme, rare climate scenarios is vital for risk and resilience models aware change, directly impacting society in different sectors. However, creating high-quality variations under-represented samples remains a challenge several generative models. This paper investigates quantizing reconstruction losses helping variational autoencoders (VAE) better synthesize extreme weather fields from conventional historical training sets. Building on the classical...