Regularization methods for the extraction of depth profiles from simulated ARXPS data derived from overlayer/substrate models

Ultrasonic waves, acquired by an array probe, are likely interfered by the crosstalk signals due to inter-communication among the elements. The crosstalk signals are not desirable and mixed with the main signals, degrading the data quality. In this work, we choose a filtering strategy to separate crosstalk and the desired signals. We design a traditional adaptive crosstalk canceller (ACC) in the Radon (intercept time, ( and slowness, p) domain. A Normalized-Least-Mean-Square strategy is used to determine the filter coefficients. Formulating the filter in the (-p space has the merits to efficiently remove random noise and significantly enhance signal-to-noise ratio. The filtered Radon panels are then transformed back to the t-x space using the inverse Radon transform. The robustness and accuracy of the filter are studied using simulated noiseless and noisy data sets. The filtered signals tracks well with the simulated noiseless signals with an average mean-square-error (MSE) of 11%. Finally the ACC is applied to an experimental data set of a 2.4-cm thick Plexiglas plate. The result has demonstrated the ACC effectively recovered the main arrivals with reasonable coherency and continuity.

We investigated the estimation of the ensemble channelized Hotelling observer (CHO) signal-to-noise ratio (SNR) for ordered-subset (OS) image reconstruction using noisy projection data. Previously, we computed the ensemble CHO SNR using a method for approximating the channelized covariance of OS reconstruction, which requires knowledge of the noise-free projection data. Here, we use a “plug-in” approach, in which noisy data is used in place of the noise-free data in the aforementioned channelized covariance approximation. Additionally, we evaluated the use of smoothing of the noisy projections before use in the covariance approximation. Additionally, we evaluated the use of smoothing of the noisy projections before use in the covariance calculation. The task was detection of a 10% contrast Gaussian signal within a slice of the MCAT phantom. Simulated projections of the MCAT phantom were scaled and Poisson noise was added to create 100 noisy signal-absent data sets. Simulated projections of the scaled signal were then added to the noisy background projections to create 100 noisy signal-present data set. These noisy data sets were then used to generate 100 estimates of the ensemble CHO SNR for reconstructions at various iterates. For comparison purposes, the same calculation was repeated with the noise-free data. The results, reported as plots of the average CHO SNR generated in this fashion, along with 95% confidence intervals, demonstrate that this approach works very well, and would allow optimization of imaging systems and reconstruction methods using a more accurate object model (i.e., real patient data).

A highly robust estimator for regression models

Simultaneously estimating the kinematic properties and the spatial extent of an object based on noisy point clouds is a ubiquitous problem in a variety of environment perception applications. For this purpose, we propose a novel approach for elliptical extended objects that employs a Kalman filter to estimate the kinematic state and a neural network architecture to infer the axes length. The key idea is to transform the point cloud measurement to a two-channel image, which serves as the input to a convolutional neural network. This input image incorporates uncertainties and is recursively updated as new measurements become available. The network is trained with simulated data using varying measurement noise levels, where measurements are assumed to be uniformly distributed across the elliptical object's surface. Noisy simulated data is used to evaluate the proposed method, and it is found to perform competitively with state-of-the-art methods. In the considered challenging scenarios with low measurement rate, it is even capable of outperforming them.

In this paper, for the first time the inverse problem of reconstructing the time-dependent potential (TDP) and displacement distribution in the hyperbolic problem with periodic boundary conditions (BCs) and nonlocal initial supplemented by over-determination measurement is numerically investigated. Though the inverse problem under consideration is ill-posed by being unstable to noise in the input data, it has a unique solution. The Crank–Nicolson-finite difference method (CN-FDM) along with the Tikhonov regularization (TR) is applied for calculating an accurate and stable numerical solution. The programming language MATLAB built-in lsqnonlin is used to solve the obtained nonlinear minimization problem. The simulated noisy input data can be inverted by both analytical and numerically simulated. The obtained results show that they are accurate and stable. The stability analysis is performed by using Fourier series.

Seismic data denoising is a critical component of seismic data processing, yet effectively removing erratic noise, characterized by its non-Gaussian distribution and high amplitude, remains a substantial challenge for conventional methods and deep-learning (DL) algorithms. Supervised learning frameworks typically outperform others, but they require pairs of noisy data sets alongside corresponding clean ground truth, which is impractical for real-world seismic data sets. In contrast, unsupervised learning (UL) methods, which do not rely on ground truth during training, often fall short in performance when compared with their supervised or traditional denoising counterparts. Moreover, current unsupervised DL methods fail to address the specific challenges posed by erratic seismic noise adequately. This paper introduces a novel zero-shot unsupervised DL framework designed specifically to mitigate random and erratic noise, with a particular emphasis on blended noise. Drawing inspiration from Noise2Noise (N2N) and data augmentation principles, we develop a robust self-supervised denoising network called robust Noiser2Noiser. Our approach eliminates the need for paired noisy and clean data sets as required by supervised methods or paired noisy data sets as in N2N. Instead, our framework relies solely on the original noisy seismic data set. Our methodology generates two independent recorrupted data sets from the original noisy data set, using one as the input and the other as the training target. Subsequently, we use a DL-based denoiser, denoising convolutional neural network, for training purposes. To address various types of random and erratic noise, the original noisy data set is recorrupted with the same noise type. Detailed explanations for generating training input and target data for blended data are provided. We apply our network to synthetic and real marine data examples, demonstrating significantly improved noise attenuation performance compared with traditional denoising methods and state-of-the-art UL Codes are available on https://github.com/Ji-seismic/N2N_deblending.

Forward and inverse problems in towed cable hydrodynamics

Nuclear resonance scattering (NRS) 16O(α,α)16O at 3.045 MeV (Γ=10 keV) has been used for oxygen depth profiling in various thin oxide films. There are two ways by which the oxygen concentration versus depth profile can be obtained from the experimental data: energy spectrum simulation or yield distribution analysis. Energy spectrum simulation is done using the standard RBS software/Rutherford Universal Manipulation Program (RUMP) where only one spectrum is usually needed from the measurement. Yield distribution analysis is accomplished by using a custom developed software/Resonance Analysis Program (RAP) and involves a series of spectra obtained by stepping up the beam energy above the resonance energy. This article aims at comparing the fundamentals of both methods and also discussing their advantages and disadvantages in terms of the data acquisition and the post data analysis. A thermally grown thick SiO2 film and a thin titanium oxide film grown by corona point discharge were examined.

Ultrasonic guided waves are gaining acceptance for structural health monitoring and nondestructive evaluation of plate-like structures. One configuration of interest is a spatially distributed array of fixed piezoelectric devices. Typical operation consists of recording signals from all transmit-receive pairs and subtracting pre-recorded baselines to detect changes, possibly due to damage or other effects. While techniques such as delay-and-sum imaging as applied to differential signals are both simple and capable of detecting flaws, their performance is limited, particularly when there are multiple damage sites. Here a very different approach to imaging is considered that exploits the expected sparsity of structural damage; i.e., the structure is mostly damage-free. Differential signals are decomposed into a sparse linear combination of location-based components, which are pre-computed from a simple propagation model. The sparse reconstruction techniques of basis pursuit denoising and orthogonal matching pursuit are applied to achieve this decomposition, and a hybrid reconstruction method is also proposed and evaluated. Noisy simulated data and experimental data recorded on an aluminum plate with artificial damage are considered. Results demonstrate the efficacy of all three methods by producing very sparse indications of damage at the correct locations even in the presence of model mismatch and significant noise.

We introduce a 4-dimensional joint generative probabilistic model for estimation of activity in a PET/MRI imaging system. The model is based on a mixture of Gaussians, relating time dependent activity and MRI image intensity to a hidden static variable, allowing one to estimate jointly activity, the parameters that capture the interdependence of the two images and motion parameters. An iterative algorithm for optimisation of the model is described. Noisy simulation data, modeling 3-D patient head movements, is obtained with realistic PET and MRI simulators and with a brain phantom from the BrainWeb database. Joint estimation of activity and motion parameters within the same framework allows us to use information from the MRI images to improve the activity estimate in terms of noise and recovery.

In this paper a unified theory is presented for system identification in structural dynamics using recursive frequency-domain filters. Output error, equation error, and instrumental variable (I. V.) filters are developed as special cases. Parameter estimates are extracted from noisy simulated data from a multi-degree-of-freedom system. The equation error filter is strongly convergent but is prone to asymptotic bias error in the presence of measurement noise. An I. V. filter is developed in order to overcome the bias problem while maintaining the desirable convergence properties of the equation error approach.

A recursive algorithm, which appears to be new, is presented for estimating the amplitude and phase of a wave field from intensity-only measurements on two or more scan planes at different axial positions. The problem is framed as a nonlinear optimization, in which the angular spectrum of the complex field model is adjusted in order to minimize the relative entropy, or Kullback-Leibler divergence, between the measured and reconstructed intensities. The most common approach to this so-called phase retrieval problem is a variation of the well-known Gerchberg-Saxton algorithm devised by Misell (J. Phys. D6, L6, 1973), which is efficient and extremely simple to implement. The new algorithm has a computational structure that is very similar to Misell's approach, despite the fundamental difference in the optimization criteria used for each. Based upon results from noisy simulated data, the new algorithm appears to be more robust than Misell's approach and to produce better results from low signal-to-noise ratio data. The convergence of the new algorithm is examined.

The paper describes the determination and depth profiling of oxygen in thin oxide films using 18O(p,α)15N nuclear reaction. The excitation function of 18O(p,α)15N nuclear reaction exhibits a resonance at 629 keV and a plateau at 730 keV with uniform cross-section. The resonance is used to determine the depth profile of oxygen in films while the plateau, to estimate its overall concentration. The resonance, characterized by a width of 2.1 keV enables high-depth resolution (~20 nm) measurements and has a probing depth of more than a micron. The paper presents depth profile measurements of oxygen in several metal oxide films (SiO2, TiO2 and HfO2) using this resonance. Possible interferences arising from 15N(p,α)12C, 19F(p,α)16O and 11B(p,α)2α nuclear reactions are also discussed. It has been shown that it can serve as a suitable alternative to 3.05 MeV 16O(α,α)16O resonant scattering which is generally used for depth profiling oxygen. It is, in fact, more reliable and precise in materials that witness significant large angle multiple scattering. It can also be advantageously used to monitor 18O, when used as a tracer.

This work presents a numerical study of the fluence-resolution behaviour for two coherent lensless X-ray imaging techniques. To this end the fluence-resolution relationship of inline near-field holography and far-field coherent diffractive imaging are compared in numerical experiments. To achieve this, the phase reconstruction is carried out using iterative phase-retrieval algorithms on simulated noisy data. Using the incident photon fluence on the specimen as the control parameter, the achievable resolution for two example phantoms (cell and bitmap) is studied. The results indicate the superior performance of holography compared with coherent diffractive imaging, for the same fluence and phase-reconstruction procedure.

Chapter 42 - A Kinetic Model for Double Injection 18FDG Studies