Showing 1–9 of 9 results for author: Boßmann, F

Improved impedance inversion by deep learning and iterated graph Laplacian

Authors: Davide Bianchi, Florian Bossmann, Wenlong Wang, Mingming Liu

Abstract: Deep learning techniques have shown significant potential in many applications through recent years. The achieved results often outperform traditional techniques. However, the quality of a neural network highly depends on the used training data. Noisy, insufficient, or biased training data leads to suboptimal results. We present a hybrid method that combines deep learning with iterated graph Lap… ▽ More Deep learning techniques have shown significant potential in many applications through recent years. The achieved results often outperform traditional techniques. However, the quality of a neural network highly depends on the used training data. Noisy, insufficient, or biased training data leads to suboptimal results. We present a hybrid method that combines deep learning with iterated graph Laplacian and show its application in acoustic impedance inversion which is a routine procedure in seismic explorations. A neural network is used to obtain a first approximation of the underlying acoustic impedance and construct a graph Laplacian matrix from this approximation. Afterwards, we use a Tikhonov-like variational method to solve the impedance inversion problem where the regularizer is based on the constructed graph Laplacian. The obtained solution can be shown to be more accurate and stable with respect to noise than the initial guess obtained by the neural network. This process can be iterated several times, each time constructing a new graph Laplacian matrix from the most recent reconstruction. The method converges after only a few iterations returning a much more accurate reconstruction. We demonstrate the potential of our method on two different datasets and under various levels of noise. We use two different neural networks that have been introduced in previous works. The experiments show that our approach improves the reconstruction quality in the presence of noise. △ Less

Submitted 25 April, 2024; originally announced April 2024.

Report number: submitted to SEG Geophysics (June 2024)

A fast and gridless ORKA algorithm for tracking moving and deforming objects

Authors: Florian Bossmann, Jianwei Ma, Wenze wu

Abstract: Identifying objects in given data is a task frequently encountered in many applications. Finding vehicles or persons in video data, tracking seismic waves in geophysical exploration data, or predicting a storm front movement from meteorological measurements are only some of the possible applications. In many cases, the object of interest changes its form or position from one measurement to another… ▽ More Identifying objects in given data is a task frequently encountered in many applications. Finding vehicles or persons in video data, tracking seismic waves in geophysical exploration data, or predicting a storm front movement from meteorological measurements are only some of the possible applications. In many cases, the object of interest changes its form or position from one measurement to another. For example, vehicles in a video may change its position or angle to the camera in each frame. Seismic waves can change its arrival time, frequency, or intensity depending on the sensor position. Storm fronts can change its form and position over time. This complicates the identification and tracking as the algorithm needs to deal with the changing object over the given measurements. In a previous work, the authors presented a new algorithm to solve this problem - Object reconstruction using K-approximation (ORKA). The algorithm can solve the problem at hand but suffers from two disadvantages. On the one hand, the reconstructed object movement is bound to a grid that depends on the data resolution. On the other hand, the complexity of the algorithm increases exponentially with the resolution. We overcome both disadvantages by introducing an iterative strategy that uses a resampling method to create multiple resolutions of the data. In each iteration the resolution is increased to reconstruct more details of the object of interest. This way, we can even go beyond the original resolution by artificially upsampling the data. We give error bounds and a complexity analysis of the new method. Furthermore, we analyze its performance in several numerical experiments as well as on real data. We also give a brief introduction on the original ORKA algorithm. Knowledge of the previous work is thus not required. △ Less

Submitted 4 February, 2024; originally announced February 2024.

Multiresolution ORKA: fast and resolution independent object reconstruction using a K-approximation graph

Authors: Florian Bossmann, Wenze Wu

Abstract: Object recognition and reconstruction is of great interest in many research fields. Detecting pedestrians or cars in traffic cameras or tracking seismic waves in geophysical exploration are only two of many applications. Recently, the authors developed a new method - Object reconstruction using K-approximation (ORKA) - to extract such objects out of given data. In this method a special object mode… ▽ More Object recognition and reconstruction is of great interest in many research fields. Detecting pedestrians or cars in traffic cameras or tracking seismic waves in geophysical exploration are only two of many applications. Recently, the authors developed a new method - Object reconstruction using K-approximation (ORKA) - to extract such objects out of given data. In this method a special object model is used where the movement and deformation of the object can be controlled to fit the application. ORKA in its current form is highly dependent on the data resolution. On the one hand, the movement of the object can only be reconstructed on a grid that depends on the data resolution. On the other hand, the runtime increases exponentially with the resolution. Hence, the resolution of the data needs to be in a small range where the reconstruction is accurate enough but the runtime is not too high. In this work, we present a multiresolution approach, where we combine ORKA with a wavelet decomposition of the data. The object is then reconstructed iteratively what drastically reduces the runtime. Moreover, we can increase the data resolution such that the movement reconstruction no longer depends on the original grid. We also give a brief introduction on the original ORKA algorithm. Hence, knowledge of the previous work is not required. △ Less

Submitted 17 November, 2022; originally announced November 2022.

ORKA: Object reconstruction using a K-approximation graph

Authors: Florian Bossmann, Jianwei Ma

Abstract: Data processing has to deal with many practical difficulties. Data is often corrupted by artifacts or noise and acquiring data can be expensive and difficult. Thus, the given data is often incomplete and inaccurate. To overcome these problems, it is often assumed that the data is sparse or low-dimensional in some domain. When multiple measurements are taken, this sparsity often appears in a struct… ▽ More Data processing has to deal with many practical difficulties. Data is often corrupted by artifacts or noise and acquiring data can be expensive and difficult. Thus, the given data is often incomplete and inaccurate. To overcome these problems, it is often assumed that the data is sparse or low-dimensional in some domain. When multiple measurements are taken, this sparsity often appears in a structured manner. We propose a new model that assumes the data only contains a few relevant objects, i.e., it is sparse in some object domain. We model an object as a structure that can only change slightly in form and continuously in position over different measurements. This can be modeled by a matrix with highly correlated columns and a column shift operator that we introduce in this work. We present an efficient algorithm to solve the object reconstruction problem based on a K-approximation graph. We prove optimal approximation bounds and perform a numerical evaluation of the method. Examples from applications including Geophysics, video processing, and others will be given. △ Less

Submitted 27 September, 2022; v1 submitted 1 June, 2022; originally announced June 2022.

Journal ref: Inverse Problems 38 125009, 2022

Structural Sparsity in Multiple Measurements

Authors: Florian Boßmann, Sara Krause-Solberg, Johannes Maly, Nada Sissouno

Abstract: We propose a novel sparsity model for distributed compressed sensing in the multiple measurement vectors (MMV) setting. Our model extends the concept of row-sparsity to allow more general types of structured sparsity arising in a variety of applications like, e.g., seismic exploration and non-destructive testing. To reconstruct structured data from observed measurements, we derive a non-convex but… ▽ More We propose a novel sparsity model for distributed compressed sensing in the multiple measurement vectors (MMV) setting. Our model extends the concept of row-sparsity to allow more general types of structured sparsity arising in a variety of applications like, e.g., seismic exploration and non-destructive testing. To reconstruct structured data from observed measurements, we derive a non-convex but well-conditioned LASSO-type functional. By exploiting the convex-concave geometry of the functional, we design a projected gradient descent algorithm and show its effectiveness in extensive numerical simulations, both on toy and real data. △ Less

Submitted 31 December, 2021; v1 submitted 2 March, 2021; originally announced March 2021.

Comments: Copyright 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

A direct solver for the phase retrieval problem in ptychographic imaging

Authors: Nada Sissouno, Florian Boßmann, Frank Filbir, Mark Iwen, Maik Kahnt, Rayan Saab, Christian Schroer, Wolfgang zu Castell

Abstract: Measurements achieved with ptychographic imaging are a special case of diffraction measurements. They are generated by illuminating small parts of a sample with, e.g., a focused X-ray beam. By shifting the sample, a set of far-field diffraction patterns of the whole sample are then obtained. From a mathematical point of view those measurements are the squared modulus of the windowed Fourier transf… ▽ More Measurements achieved with ptychographic imaging are a special case of diffraction measurements. They are generated by illuminating small parts of a sample with, e.g., a focused X-ray beam. By shifting the sample, a set of far-field diffraction patterns of the whole sample are then obtained. From a mathematical point of view those measurements are the squared modulus of the windowed Fourier transform of the sample. Thus, we have a phase retrieval problem for local Fourier measurements. A direct solver for this problem was introduced by Iwen, Viswanathan and Wang in 2016 and improved by Iwen, Preskitt, Saab and Viswanathan in 2018. Motivated by the applied perspective of ptychographic imaging, we present a generalization of this method and compare the different versions in numerical experiments. The new method proposed herein turns out to be more stable, particularly in the case of missing data. △ Less

Submitted 16 April, 2019; originally announced April 2019.

Enhanced image approximation using shifted rank-1 reconstruction

Authors: Florian Boßmann, Jianwei Ma

Abstract: Low rank approximation has been extensively studied in the past. It is most suitable to reproduce rectangular like structures in the data. In this work we introduce a generalization using shifted rank-1 matrices to approximate $A\in\mathbb{C}^{M\times N}$. These matrices are of the form $S_λ(uv^*)$ where $u\in\mathbb{C}^M$, $v\in\mathbb{C}^N$ and $λ\in\mathbb{Z}^N$.The operator $S_λ$ circularly sh… ▽ More Low rank approximation has been extensively studied in the past. It is most suitable to reproduce rectangular like structures in the data. In this work we introduce a generalization using shifted rank-1 matrices to approximate $A\in\mathbb{C}^{M\times N}$. These matrices are of the form $S_λ(uv^*)$ where $u\in\mathbb{C}^M$, $v\in\mathbb{C}^N$ and $λ\in\mathbb{Z}^N$.The operator $S_λ$ circularly shifts the k-th column of $uv^*$ by $λ_k$. These kind of shifts naturally appear in applications, where an object $u$ is observed in $N$ measurements at different positions indicated by the shift $λ$. The vector $v$ gives the observation intensity. Exemplary, a seismic wave can be recorded at $N$ sensors with different time of arrival $λ$; Or a car moves through a video changing its position in every frame. We present theoretical results as well as an efficient algorithm to calculate a shifted rank-1 approximation in $O(NM \log M)$. The benefit of the proposed method is demonstrated in numerical experiments. A comparison to other sparse approximation methods is given. Finally, we illustrate the utility of the extracted parameters for direct information extraction in several applications including video processing or non-destructive testing. △ Less

Submitted 3 October, 2018; originally announced October 2018.

Modeling variational inpainting methods with splines

Authors: Florian Boßmann, Tomas Sauer, Nada Sissouno

Abstract: Mathematical methods of image inpainting involve the discretization of given continuous models. We present a method that avoids the standard pointwise discretization by modeling known variational approaches, in particular total variation (TV), using a finite dimensional spline space. Besides the analysis of the resulting model, we present a numerical implementation based on the alternating method… ▽ More Mathematical methods of image inpainting involve the discretization of given continuous models. We present a method that avoids the standard pointwise discretization by modeling known variational approaches, in particular total variation (TV), using a finite dimensional spline space. Besides the analysis of the resulting model, we present a numerical implementation based on the alternating method of multipliers. We compare the results numerically with classical TV inpainting and give examples of applications. △ Less

Submitted 23 May, 2017; originally announced May 2017.

Comments: 16 pages, 10 figures

MSC Class: 41A15; 41A63; 49M25; 65D05; 65D07; 65K10

Generalized orthogonal matching pursuit for multiple measurements - A structural approach

Authors: Florian Boßmann

Abstract: Sparse data approximation has become a popular research topic in signal processing. However, in most cases only a single measurement vector (SMV) is considered. In applications, the multiple measurement vector (MMV) case is more usual, i.e., the sparse approximation problem has to be solved for several data vectors coming from closely related measurements. Thus, there is an unknown inter-vector co… ▽ More Sparse data approximation has become a popular research topic in signal processing. However, in most cases only a single measurement vector (SMV) is considered. In applications, the multiple measurement vector (MMV) case is more usual, i.e., the sparse approximation problem has to be solved for several data vectors coming from closely related measurements. Thus, there is an unknown inter-vector correlation between the data vectors. Using SMV methods typically does not return the best approximation result as the correlation is ignored. In the past few years several algorithms for the MMV case have been designed to overcome this problem. Most of these techniques focus on the approximation quality while quite strong assumptions to the type of inter-vector correlation are made. While we still want to find a sparse approximation, our focus lies on preserving (possibly complex) structures in the data. Structural knowledge is of interest in many applications. It can give information about e.g., type, form, number or size of objects of interest. This may even be more useful than information given by the non-zero amplitudes itself. Moreover, it allows efficient post processing of the data. We numerically compare our new approach with other techniques and demonstrate its benefits in two applications. △ Less

Submitted 23 May, 2017; originally announced May 2017.

Comments: 10 pages, 9 figures

MSC Class: 41A29; 65D99 ACM Class: G.1.2