A5100364756- Geophysical and Geoelectrical Methods
- Music and Audio Processing
- Speech and Audio Processing
- Video Coding and Compression Technologies
- Blind Source Separation Techniques
- Underwater Acoustics Research
- Advanced Data Compression Techniques
- Geophysical Methods and Applications
- Image and Signal Denoising Methods
- Nonlinear Differential Equations Analysis
- Seismic Imaging and Inversion Techniques
- Speech Recognition and Synthesis
- Non-Destructive Testing Techniques
- Seismic Waves and Analysis
- Music Technology and Sound Studies
- Remote Sensing and Land Use
- Paleontology and Stratigraphy of Fossils
- Image and Video Quality Assessment
- Geological and Geochemical Analysis
- Advanced Vision and Imaging
- Geochemistry and Elemental Analysis
- Geology and Paleoclimatology Research
- Nonlinear Partial Differential Equations
- Neural Networks and Applications
- High-pressure geophysics and materials
Tianjin University
2025
Hunan Normal University
2007-2024
Changsha Normal University
2022-2024
Central South University
2011-2024
Shanghai Medical College of Fudan University
2024
Children's Hospital of Fudan University
2024
Air Force Engineering University
2022-2024
Peking University
2024
Chinese Academy of Geological Sciences
2010-2023
Hefei University of Technology
2013-2023
Audio magnetotellurics (AMT), as a commonly used passive geophysical technique, provides outstanding metal ore exploration capabilities based on the resistivity structure of earth. However, accuracy AMT in translating geoelectrical structures decreases when data collected mining areas are poor quality and contain complex anthropogenic noise, leading to distorted apparent resistivity-phase curves posing significant challenges for mineral exploration. To effectively denoise data, we develop...
It is well known that a magnetotelluric (MT) signal with high signal-to-noise ratio an important prerequisite for correct interpretation of subsurface structures. However, MT signals collected in the environment strong cultural noise often are low data quality due to pollution, which seriously affects accuracy interpretation. As can be seen from time-domain waveform, highly energetic, diverse, and random. This means denoising methods should have applicability guarantee accurate effective...
ABSTRACT Magnetotelluric is one of the mainstream exploration geophysical methods, which plays a vital role in studying deep geological structures and finding hidden blind ore bodies. The seriousness human electromagnetic noise causes large number abnormal waveforms time series measured magnetotelluric data, data can no longer objectively reflect underground electrical distribution. In this work, we propose processing method based on K singular value decomposition dictionary training. First,...
The magnetotelluric (MT) data collected in an ore-concentration area are extremely vulnerable to all kinds of noise pollution. However, separating real MT signals from strong is still a difficult problem, and the quite distinct clean morphological features. By performing signal-noise identification prediction, we develop deep learning method denoise containing noise. First, use convolutional neural network (CNN) learn feature differences between samples massive learned features realize...
Abstract The ocean is hypothesized to have been anoxic throughout the Marinoan “Snowball Earth” event, from ca. 649 635 Ma, with potentially catastrophic implications for survival of eukaryotic life. However, precise nature redox chemistry across this critical interval, and hence factors that governed persistence eukaryotes, remains unknown. We report records pyrite iron sulfur isotopes, combined Fe speciation, glaciogenic diamictites Nantuo Formation South China. These data provide...
To avoid the blindness of overall de-noising method and retain useful low frequency signals that are not over processed, we proposed a novel audio magnetotelluric (AMT) signal-noise identification separation based on multifractal spectrum matching pursuit. We extracted two sets characteristic from AMT time-series data to analyze singularity. used support vector machine approach learn characteristics in sample’s library generate model distinguish between sections with without interference...
In the field of adversarial attacks, generative network (GAN) has shown better performance. There have been few studies applying it to malware sample supplementation, due complexity handling discrete data. More importantly, unbalanced family samples interfere with analytical power detection models and mislead classification. To address problem impact imbalance on accuracy, a selection feature conditional Wasserstein (SFCWGAN) bidirectional temporal convolutional (BiTCN) are proposed. First,...
Abstract Magnetotelluric (MT) data processing can increase the reliability of measured data. Traditional MT denoising methods are usually applied to entire time-series, which results in loss useful signals and a decrease imaging accuracy electromagnetic inversion. However, targeted noise separation retain part signal unaffected by strong enhance quality responses. Thus, we propose novel method for that uses refined composite multiscale dispersion entropy (RCMDE) orthogonal matching pursuit...
Titanium isotopes are emerging as a power tool for studying magmatic processes on the Earth and other planets. In our work, novel robust method in situ Ti isotopic analysis of titanium-bearing minerals was presented by fs-LA-MC-ICP-MS.
As magnetotelluric (MT) is an important method for exploring the geoelectrical structure of underground, it has motivated in-depth research and application by many geophysicists. Nevertheless, due to influence environment, collected data are interfered with strong humanistic noise, which might result in a loss their authenticity. To solve above problems, time-frequency domain methods have emerged. Based on advantages singular value decomposition (SVD) denoising, we propose noisy processing...
Geometry partitioning for video coding involves establishing a partition line boundary within each block-shaped region and applying motion-compensated prediction to the two sub-regions created by line. This paper presents techniques enhancing effectiveness reducing complexity of geometry schemes. A texture-difference-based approach is described simplify process selecting lines. Applying this together with skipping strategy blocks uniform texture can achieve 94% reduction encoding time while...