arXiv subjects

Florian Römer

Publications and source records attributed to Florian Römer.

10 recordsLinked to original sources

arXiv2026

Compressed BC-LISTA via Low-Rank Convolutional Decomposition

We study Sparse Signal Recovery (SSR) methods for multichannel imaging with compressed {forward and backward} operators that preserve reconstruction accuracy. We propose a Compressed Block-Convolutional (C-BC) measurement model based on a low-rank Convolutional Neural Network (CNN) decomposition that is analytically initialized from a low-rank factorization of physics-derived forward/backward operators in time delay-based measurements. We use Orthogonal Matching Pursuit (OMP) to select a compact set of basis filters from the analytic model and compute linear mixing coefficients to approximate the full model. We consider the Learned Iterative Shrinkage-Thresholding Algorithm (LISTA) network as a representative example for which the C-BC-LISTA extension is presented. In simulated multichannel ultrasound imaging across multiple Signal-to-Noise Ratios (SNRs), C-BC-LISTA requires substantially fewer parameters and smaller model size than other state-of-the-art (SOTA) methods while improving reconstruction accuracy. In ablations over OMP, Singular Value Decomposition (SVD)-based, and random initializations, OMP-initialized structured compression performs best, yielding the most efficient training and the best performance.

eess.IV↗

arXiv2025

Learning Structured Compressed Sensing with Automatic Resource Allocation

Multidimensional data acquisition often requires extensive time and poses significant challenges for hardware and software regarding data storage and processing. Rather than designing a single compression matrix as in conventional compressed sensing, structured compressed sensing yields dimension-specific compression matrices, reducing the number of optimizable parameters. Recent advances in machine learning (ML) have enabled task-based supervised learning of subsampling matrices, albeit at the expense of complex downstream models. Additionally, the sampling resource allocation across dimensions is often determined in advance through heuristics. To address these challenges, we introduce Structured COmpressed Sensing with Automatic Resource Allocation (SCOSARA) with an information theory-based unsupervised learning strategy. SCOSARA adaptively distributes samples across sampling dimensions while maximizing Fisher information content. Using ultrasound localization as a case study, we compare SCOSARA to state-of-the-art ML-based and greedy search algorithms. Simulation results demonstrate that SCOSARA can produce high-quality subsampling matrices that achieve lower Cramér-Rao Bound values than the baselines. In addition, SCOSARA outperforms other ML-based algorithms in terms of the number of trainable parameters, computational complexity, and memory requirements while automatically choosing the number of samples per axis.

cs.LG↗

arXiv2024

Misspecification of Multiple Scattering in Scalar Wave Fields and its Impact in Ultrasound Tomography

In this work, we investigate the localization of targets in the presence of multiple scattering. We focus on the often omitted scenario in which measurement data is affected by multiple scattering, and a simpler model is employed in the estimation. We study the impact of such model mismatch by means of the Misspecified Cramér-Rao Bound (MCRB). In numerical simulations inspired by tomographic inspection in ultrasound nondestructive testing, the MCRB is shown to correctly describe the estimation variance of localization parameters under misspecification of the wave propagation model. We provide extensive discussion on the utility of the MCRB in the practical task of verifying whether a chosen misspecified model is suitable for localization based on the properties of the maximum likelihood estimator and the nuanced distinction between bias and parameter space differences. Finally, we highlight that careful interpretation is needed whenever employing the classical CRB in the presence of mismatch through numerical examples based on the Born approximation and other simplified propagation models stemming from it.

eess.SP↗

arXiv2024

Efficient Convolutional Forward Modeling and Sparse Coding in Multichannel Imaging

This study considers the Block-Toeplitz structural properties inherent in traditional multichannel forward model matrices, using Full Matrix Capture (FMC) in ultrasonic testing as a case study. We propose an analytical convolutional forward model that transforms reflectivity maps into FMC data. Our findings demonstrate that the convolutional model excels over its matrix-based counterpart in terms of computational efficiency and storage requirements. This accelerated forward modeling approach holds significant potential for various inverse problems, notably enhancing Sparse Signal Recovery (SSR) within the context LASSO regression, which facilitates efficient Convolutional Sparse Coding (CSC) algorithms. Additionally, we explore the integration of Convolutional Neural Networks (CNNs) for the forward model, employing deep unfolding to implement the Learned Block Convolutional ISTA (BC-LISTA).

eess.SP↗

arXiv2024

Integrating AI in NDE: Techniques, Trends, and Further Directions

The digital transformation is fundamentally changing our industries, affecting planning, execution as well as monitoring of production processes in a wide range of application fields. With product line-ups becoming more and more versatile and diverse, the necessary inspection and monitoring sparks significant novel requirements on the corresponding Nondestructive Evaluation (NDE) systems. The establishment of increasingly powerful approaches to incorporate Artificial Intelligence (AI) may provide just the needed innovation to solve some of these challenges. In this paper we provide a comprehensive survey about the usage of AI methods in NDE in light of the recent innovations towards NDE 4.0. Since we cannot discuss each NDE modality in one paper, we limit our attention to magnetic methods, ultrasound, thermography, as well as optical inspection. In addition to reviewing recent AI developments in each field, we draw common connections by pointing out NDE-related tasks that have a common underlying mathematical problem and categorizing the state of the art according to the corresponding sub-tasks. In so doing, interdisciplinary connections are drawn that provide a more complete overall picture.

eess.SP↗

arXiv2020

Frequency Sub-Sampling of Ultrasound Non-Destructive Measurements: Acquisition, Reconstruction and Performance

In ultrasound nondestructive testing, a widespread approach is to take synthetic aperture measurements from the surface of a specimen to detect and locate defects within it. Based on these measurements, imaging is usually performed using the Synthetic Aperture Focusing Technique (SAFT). However, SAFT is sub-optimal in terms of resolution and requires oversampling in time domain to obtain a fine grid for the Delay-and-Sum (DAS). On the other hand, parametric reconstruction algorithms give better resolution, but their usage for imaging becomes computationally expensive due to the size of the parameter space and the large amount of measurement data in realistic 3-D scenarios. In the literature, the remedies to this are twofold: First, the amount of measurement data can be reduced using state of the art sub-Nyquist sampling approaches to measure Fourier coefficients instead of time domain samples. Second, parametric reconstruction algorithms mostly rely on matrix-vector operations that can be implemented efficiently by exploiting the underlying model structure. In this paper, we propose and compare different strategies to choose the Fourier coefficients to be measured. Their asymptotic performance is compared by numerically evaluating the Cramér-Rao-Bound for the localizability of the defect coordinates. These subsampling strategies are then combined with an $\ell_1$-minimization scheme to compute 3-D reconstructions from the low-rate measurements. Compared to conventional DAS, this allows us to formulate a fully physically motivated forward model. To enable this, the projection operations of the forward model matrix are implemented matrix-free by exploiting the underlying 2-level Toeplitz structure. Finally, we show that high resolution reconstructions from as low as a single Fourier coefficient per scan are possible based on simulated data as well as on measurements.

eess.SP↗

arXiv2018

Combining Matrix Design for 2D DoA Estimation with Compressive Antenna Arrays using Stochastic Gradient Descent

Recently, compressive antenna arrays have been considered for DoA estimation with reduced hardware complexity. By utilizing compressive sensing, such arrays employ a linear combining network to combine signals from a larger set of antenna elements in the analog RF domain. In this paper, we develop a design approach based on the minimization of error between spatial correlation function (SCF) of the compressive and the uncompressed array resulting in the estimation performance of the two arrays to be as close as possible. The proposed design is based on grid-free stochastic gradient descent (SGD) optimization. In addition to a low computational cost for the proposed method, we show numerically that the resulting combining matrices perform better than the ones generated by a previous approach and combining matrices generated from a Gaussian ensemble.

eess.SP↗

arXiv2018

ADMM for ND Line Spectral Estimation using Grid-Free Compressive Sensing from Multiple Measurements with Applications to DOA Estimation

This paper is concerned with estimating unknown multi-dimensional frequencies from linear compressive measurements. This is accomplished by employing the recently proposed atomic norm minimization framework to recover these frequencies under a sparsity prior without imposing any grid restriction on these frequencies. To this end, we give a rigorous derivation of an iterative scheme called alternating direction of multipliers method, which is able to incorporate multiple compressive snapshots from a multi-dimensional superposition of complex harmonics. The key result here is how to formulate the objective function minimized by this scheme and its partial derivatives, which become hard to manage if the dimensionality of the frequencies is larger than 1. Moreover we demonstrate the performance of this approach in case of 3D line spectral estimation and 2D DOA estimation with a synthetic antenna array.

eess.SP↗

arXiv2017

Sparsity Order Estimation from a Single Compressed Observation Vector

We investigate the problem of estimating the unknown degree of sparsity from compressive measurements without the need to carry out a sparse recovery step. While the sparsity order can be directly inferred from the effective rank of the observation matrix in the multiple snapshot case, this appears to be impossible in the more challenging single snapshot case. We show that specially designed measurement matrices allow to rearrange the measurement vector into a matrix such that its effective rank coincides with the effective sparsity order. In fact, we prove that matrices which are composed of a Khatri-Rao product of smaller matrices generate measurements that allow to infer the sparsity order. Moreover, if some samples are used more than once, one of the matrices needs to be Vandermonde. These structural constraints reduce the degrees of freedom in choosing the measurement matrix which may incur in a degradation in the achievable coherence. We thus also address suitable choices of the measurement matrices. In particular, we analyze Khatri-Rao and Vandermonde matrices in terms of their coherence and provide a new design for Vandermonde matrices that achieves a low coherence.

cs.IT↗

arXiv2016

Design and Analysis of Compressive Antenna Arrays for Direction of Arrival Estimation

In this paper we investigate the design of compressive antenna arrays for direction of arrival (DOA) estimation that aim to provide a larger aperture with a reduced hardware complexity by a linear combination of the antenna outputs to a lower number of receiver channels. We present a basic receiver architecture of such a compressive array and introduce a generic system model that includes different options for the hardware implementation. We then discuss the design of the analog combining network that performs the receiver channel reduction, and propose two design approaches. The first approach is based on the spatial correlation function which is a low-complexity scheme that in certain cases admits a closed-form solution. The second approach is based on minimizing the Cramer-Rao Bound (CRB) with the constraint to limit the probability of false detection of paths to a pre-specified level. Our numerical simulations demonstrate the superiority of the proposed optimized compressive arrays compared to the sparse arrays of the same complexity and to compressive arrays with randomly chosen combining kernels.

cs.IT↗