Article
Geochemistry & Geophysics
Xinming Wu, Shangsheng Yan, Zhengfa Bi, Sibo Zhang, Hongjie Si
Summary: The article proposes an improved deep learning method by replacing a 1D neural network with a 2D convolutional neural network and incorporating constraints of an initial impedance model. This method provides low-frequency trend control, helpful for stable impedance predictions.
Article
Geochemistry & Geophysics
Xujia Shang, Pan Zhang, Liguo Han, Yuanyun Yang, Yixiu Zhou
Summary: Traditional full waveform inversion (FWI) relies heavily on low-frequency data or accurate initial models. Passive seismic data, which contain abundant low-frequency components, can be effectively used for FWI through virtual source data obtained by seismic interferometry (SI). Inhomogeneous distribution of passive sources in the subsurface leads to artifacts in the reconstruction results using cross-correlation based SI. In this study, an improved SI method based on linear Radon transform-based MDD (LRTMDD) is proposed to mitigate this issue and achieve better reconstruction results.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Multidisciplinary Sciences
Rejwana Tasnim Rimi, K. M. Azharul Hasan, Tatsuo Tsuji
Summary: We propose an in-memory multidimensional query processing algorithm for dense data using a higher-dimensional array. The algorithm utilizes a Converted two-dimensional Array (C2A) and shows improved performance for data retrieval by optimizing data locality and cache miss rate.
SCIENTIFIC REPORTS
(2023)
Article
Mathematics
Valery Karachik
Summary: In this work, the method of using Green's function is extended to the Dirichlet boundary value problem for the polyharmonic equation, finding the harmonic components of the solution without invoking Green's function. The explicit representation of the harmonic components for the Neumann boundary value problem for the polyharmonic equation in the unit ball is obtained using solutions to the Laplace equation.
Article
Chemistry, Multidisciplinary
Joost van der Neut, Joeri Brackenhoff, Giovanni Angelo Meles, Evert Slob, Kees Wapenaar
Summary: By solving the Marchenko equation, Green's functions at an arbitrary depth level inside an unknown elastic layered medium can be retrieved from single-sided reflection data. We introduce an alternative Marchenko equation and an auxiliary equation, as well as a coupled equation based on reflection and transmission data. When broadband reflection and transmission data are available, our method can successfully invert the system and retrieve elastodynamic Green's functions from two-sided data.
APPLIED SCIENCES-BASEL
(2022)
Article
Mathematics, Applied
H. A. Hassan
Summary: This paper focuses on establishing two-dimensional sampling theorems for discrete transforms based on second order partial difference equations. It introduces a discrete type partial difference operator, investigates its spectral properties, constructs Green's function, defines kernels for orthonormal basis of eigenvectors, and proves two sampling theorems of Lagrange interpolation type. The theory presented can be extended to higher order settings.
ADVANCES IN DIFFERENCE EQUATIONS
(2021)
Article
Geochemistry & Geophysics
Zexin Wang, Han Yue
Summary: This study proposes an inversion technique that utilizes water-reverberation phases in teleseismic P coda waves to discriminate different types of earthquake slips and enable rapid inversions. Through the application of this algorithm to seven well-studied earthquakes, it is found that the EGF-based inversion technique is more effective in discriminating these earthquakes compared to inversions performed with theoretical Green's functions. The study also analyzes the potential for applying this technique to tsunami warning purposes and suggests it as a promising algorithm for efficient tsunami early warning.
JOURNAL OF GEOPHYSICAL RESEARCH-SOLID EARTH
(2022)
Article
Computer Science, Artificial Intelligence
Yang Zhao, Xiaohan Yu, Yongsheng Gao, Chunhua Shen
Summary: This study introduces a novel identity-guided human region segmentation method for person retrieval, which predicts informative region segments by learning a set of distinct region bases and grouping intermediate feature vectors based on their similarity. HRS learns region segmentation using only identity labels, making it a practical solution to person retrieval. By jointly learning global appearance and local granularity cues, HRS enables comprehensive feature representation learning and demonstrates superior performance over existing region-based methods.
PATTERN RECOGNITION
(2022)
Article
Engineering, Multidisciplinary
Shaymaa Mustafa, Arifah Bahar, Ahmad Razin Zainal Abidin, Zainal Abdul Aziz, Mohamad Darwish
Summary: This study developed three-dimensional analytical solutions for contaminant transport induced by a pumping well near a stream using Green's function approach. Results showed that parameters such as stream width, leakance coefficient, distance from river edge to well, and pumping rate significantly affected contaminant concentration. The finite width model provided a more realistic estimate for contaminant distribution, with higher contamination levels near the stream edge.
ALEXANDRIA ENGINEERING JOURNAL
(2021)
Article
Chemistry, Analytical
Xin Gu, Yinghua Shen, Chaohui Lv
Summary: With the growth of the internet, video usage has expanded in recent years. Adding suitable music to videos has become an artistic task. In this study, a method to recommend background music for videos is proposed, considering both content and emotional information.
Article
Computer Science, Artificial Intelligence
Hongyu Sun, Laurent Demanet
Summary: Passive seismic interferometry is a blind deconvolution problem that can convert noise to signal and has various applications. Traditional methods often fail to meet the requirements in realistic situations. This paper proposes using deep neural networks to solve the relationship between correlograms and empirical Green's functions, with validation on synthetic data.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Alexei Novikov, Stephen White
Summary: In this work, we introduce MISTR (Multidimensional Intersection Sparse supporT Recovery), an algorithm that utilizes the structure of multi-dimensional signals to recover the support from magnitude-only measurements with the same accuracy as the best one-dimensional algorithms. Theoretical analysis shows that MISTR can correctly recover the support of signals distributed as a Gaussian point process with high probability under certain sparsity constraints, and provides a thresholding scheme for handling noisy measurements. The algorithm's effectiveness is further demonstrated through detailed numerical experiments, showing near-linear time complexity in practice.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Computer Science, Artificial Intelligence
Zhao Huang, Wei Zhao
Summary: This study developed a neural matching network based on multidimensional service representations to improve the performance of web service discovery. By generating multidimensional representations of keywords through various methods and calculating the cosine similarity between keywords, a multidimensional similarity matrix was constructed, ultimately achieving accurate service discovery.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Mathematics
Yakir Aharonov, Jussi Behrndt, Fabrizio Colombo, Peter Schlosser
Summary: In this paper, we study the time dependent Schrodinger equation with all possible self-adjoint singular interactions located at the origin, including delta and delta'-potentials, as well as Dirichlet, Neumann, and Robin type boundary conditions. An explicit representation of the time dependent Green's function is derived, and a rigorous mathematical meaning is given to the corresponding integral for holomorphic initial conditions using Fresnel integrals. Superoscillatory functions are treated as holomorphic entire functions and are studied in the context of weak measurements in quantum mechanics.
JOURNAL OF DIFFERENTIAL EQUATIONS
(2021)
Article
Environmental Sciences
Arno Keppens, Steven Compernolle, Daan Hubert, Tijl Verhoelst, Jose Granville, Jean-Christopher Lambert
Summary: A method is developed to remove a priori information from remotely sensed atmospheric state profiles. This method utilizes Wiener deconvolution and an iterative process to obtain profile-specific deconvolution matrices. The resulting prior-free atmospheric state representations are achieved by asserting that the deconvoluted averaging kernel matrix should equal the unit matrix. The method is successfully applied to ozone profile retrievals, producing accurate results after spatiotemporal averaging.
Article
Geochemistry & Geophysics
Joeri Brackenhoff, Jan Thorbecke, Giovanni Meles, Victor Koehne, Diego Barrera, Kees Wapenaar
Summary: The text discusses implementing 3D Marchenko equations to retrieve responses to virtual sources inside the subsurface, and applying floating point compression to reduce the volume and loading time of reflection data. By using this method, accurate Green's functions are extracted for imaging without simulating a large number of virtual source points.
GEOPHYSICAL PROSPECTING
(2022)
Article
Geochemistry & Geophysics
Kees Wapenaar, Sjoerd de Ridder
Summary: The propagator matrix and Marchenko focusing function are closely related concepts that can be transferred and learned from each other, enhancing the capabilities in handling subsurface models.
Article
Geochemistry & Geophysics
Joeri Brackenhoff, Jan Thorbecke, Kees Wapenaar
Summary: In this study, virtual sources and receivers in a 3-D subsurface were created using the single-sided homogeneous Green's function representation. By applying the representation to numerical data, it was demonstrated that the approach is an improvement over the classical representation in 3-D. The effectiveness of the method was validated through experiments, showing its potential applications under certain conditions.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Acoustics
Joost van der Neut, Joeri Brackenhoff, Giovanni Meles, Lele Zhang, Evert Slob, Kees Wapenaar
Summary: By incorporating transmission data, an auxiliary equation for the forward-scattered components of the initial focusing function can be derived, which allows for including forward-scattered waveforms in Green's function estimates. This approach has been successfully demonstrated in an acoustic medium with mass density contrast and constant propagation velocity through solving both the auxiliary and Marchenko equations successively.
IEEE TRANSACTIONS ON ULTRASONICS FERROELECTRICS AND FREQUENCY CONTROL
(2022)
Article
Chemistry, Multidisciplinary
Joost van der Neut, Joeri Brackenhoff, Giovanni Angelo Meles, Evert Slob, Kees Wapenaar
Summary: By solving the Marchenko equation, Green's functions at an arbitrary depth level inside an unknown elastic layered medium can be retrieved from single-sided reflection data. We introduce an alternative Marchenko equation and an auxiliary equation, as well as a coupled equation based on reflection and transmission data. When broadband reflection and transmission data are available, our method can successfully invert the system and retrieve elastodynamic Green's functions from two-sided data.
APPLIED SCIENCES-BASEL
(2022)
Article
Acoustics
Leon Diekmann, Ivan Vasconcelos, Kees Wapenaar, Evert Slob, Roel Snieder
Summary: Marchenko-type integrals relate focusing functions and Green's functions through reflection response measured on the open surface of the volume of interest. The traditional method is limited in including evanescent, refracted, and diving waves due to wavefield decomposition and a truncated medium state.
Article
Multidisciplinary Sciences
Kees Wapenaar, Evert Slob
Summary: In this article, reciprocity and representation theorems for 3D inhomogeneous PT-symmetric materials are discussed, with some potential applications such as interferometric Green's matrix retrieval and Marchenko-type Green's matrix retrieval in PT-symmetric materials.
Article
Geochemistry & Geophysics
Aydin Shoja, Joost van der Neut, Kees Wapenaar
Summary: Geophysicists have used least-squares reverse-time migration (LSRTM) to obtain high-resolution subsurface images. A target-oriented approach to LSRTM has been proposed, which focuses the wavefield above the target of interest and reduces computational burden. However, it still requires an accurate velocity model and suffers from multiple reflections.
GEOPHYSICAL JOURNAL INTERNATIONAL
(2023)
Article
Geochemistry & Geophysics
Johno van IJsseldijk, Joost van der Neut, Jan Thorbecke, Kees Wapenaar
Summary: Geophysical monitoring of subsurface reservoirs relies on detecting small changes in seismic response using the Marchenko method to isolate the target response and reveal target-related multiples that probe the reservoir more than once. By using this method, more precise time-lapse changes can be extracted, and dynamic changes in the subsurface can be observed with increased accuracy.
Article
Geochemistry & Geophysics
Kees Wapenaar, Marcin Dukalski, Christian Reinicke, Roel Snieder
Summary: Many seismic imaging methods use wavefield extrapolation operators to redatum sources and receivers from the surface into the subsurface. We discuss wavefield extrapolation operators that account for internal multiple reflections, including propagator matrices, transfer matrices, and Marchenko focusing functions. These operators allow for accurate wavefield propagation and focusing in heterogeneous media, avoiding common approximations used in previous methods. Understanding the relationships between these operators can lead to new developments in Marchenko theory and improve applications in wavefield focusing, Green's function retrieval and imaging.
GEOPHYSICAL JOURNAL INTERNATIONAL
(2023)
Article
Geochemistry & Geophysics
Iris Hartstra, Kees Wapenaar
Summary: This study analyzes the impact of scattering on the performance of seismic interferometry applications for retrieving body-wave reflections. The results show a trade-off between the quality of the retrieved virtual primary reflection and the scattering strength of the overburden. The full-field MDD method proves to be the most resilient to the negative effects of multiple scattering.
GEOPHYSICAL JOURNAL INTERNATIONAL
(2023)
Correction
Acoustics
Kees Wapenaar
JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA
(2023)
Article
Multidisciplinary Sciences
Johno van IJsseldijk, Hadi Hajibeygi, Kees Wapenaar
Summary: Reservoir simulations for subsurface processes are crucial for geothermal energy extraction and fluid storage. Integrating simulation results with seismic surveys is challenging due to the dynamics of the reservoir altering seismic parameters. This study develops a coupled simulation and seismic methodology to better understand fluid flow in subsurface reservoirs.
SCIENTIFIC REPORTS
(2023)
Article
Geochemistry & Geophysics
Aydin Shoja, Joost van der Neut, Kees Wapenaar
Summary: Least-squares reverse-time migration (LSRTM) is a computationally demanding method used by seismologists to compute high-resolution subsurface images. This article proposes a target-enclosed LSRTM algorithm by confining the region of interest (ROI) and deriving representations for the wavefields at the upper and lower boundaries using the acoustic reciprocity theorem. The algorithm aims to compute a high-resolution image of the ROI while considering wavefields entering the target region from the surrounding boundaries. The article also investigates the potential use of virtual receivers created by Marchenko redatuming.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Johno van IJsseldijk, Joeri Brackenhoff, Jan Thorbecke, Kees Wapenaar
Summary: The data-driven Marchenko method can redatum wavefields to isolate zones of interest, resulting in clear subsurface reservoir response. It can be used in time-lapse studies to improve comparison and retrieval of characteristics. The method can also estimate internal multiples and find time-lapse traveltime differences.
GEOPHYSICAL PROSPECTING
(2023)
