Showing 1–3 of 3 results for author: Genzel, M

Near-Exact Recovery for Tomographic Inverse Problems via Deep Learning

Authors: Martin Genzel, Ingo Gühring, Jan Macdonald, Maximilian März

Abstract: This work is concerned with the following fundamental question in scientific machine learning: Can deep-learning-based methods solve noise-free inverse problems to near-perfect accuracy? Positive evidence is provided for the first time, focusing on a prototypical computed tomography (CT) setup. We demonstrate that an iterative end-to-end network scheme enables reconstructions close to numerical pr… ▽ More This work is concerned with the following fundamental question in scientific machine learning: Can deep-learning-based methods solve noise-free inverse problems to near-perfect accuracy? Positive evidence is provided for the first time, focusing on a prototypical computed tomography (CT) setup. We demonstrate that an iterative end-to-end network scheme enables reconstructions close to numerical precision, comparable to classical compressed sensing strategies. Our results build on our winning submission to the recent AAPM DL-Sparse-View CT Challenge. Its goal was to identify the state-of-the-art in solving the sparse-view CT inverse problem with data-driven techniques. A specific difficulty of the challenge setup was that the precise forward model remained unknown to the participants. Therefore, a key feature of our approach was to initially estimate the unknown fanbeam geometry in a data-driven calibration step. Apart from an in-depth analysis of our methodology, we also demonstrate its state-of-the-art performance on the open-access real-world dataset LoDoPaB CT. △ Less

Submitted 14 June, 2022; originally announced June 2022.

Comments: ICML 2022 (long talk). Code available at https://github.com/jmaces/aapm-ct-challenge. arXiv admin note: text overlap with arXiv:2106.00280

Journal ref: Proceedings of the 39th International Conference on Machine Learning, PMLR 162:7368-7381, 2022

Gradient-Based Learning of Discrete Structured Measurement Operators for Signal Recovery

Authors: Jonathan Sauder, Martin Genzel, Peter Jung

Abstract: Countless signal processing applications include the reconstruction of signals from few indirect linear measurements. The design of effective measurement operators is typically constrained by the underlying hardware and physics, posing a challenging and often even discrete optimization task. While the potential of gradient-based learning via the unrolling of iterative recovery algorithms has been… ▽ More Countless signal processing applications include the reconstruction of signals from few indirect linear measurements. The design of effective measurement operators is typically constrained by the underlying hardware and physics, posing a challenging and often even discrete optimization task. While the potential of gradient-based learning via the unrolling of iterative recovery algorithms has been demonstrated, it has remained unclear how to leverage this technique when the set of admissible measurement operators is structured and discrete. We tackle this problem by combining unrolled optimization with Gumbel reparametrizations, which enable the computation of low-variance gradient estimates of categorical random variables. Our approach is formalized by GLODISMO (Gradient-based Learning of DIscrete Structured Measurement Operators). This novel method is easy-to-implement, computationally efficient, and extendable due to its compatibility with automatic differentiation. We empirically demonstrate the performance and flexibility of GLODISMO in several prototypical signal recovery applications, verifying that the learned measurement matrices outperform conventional designs based on randomization as well as discrete optimization baselines. △ Less

Submitted 10 November, 2022; v1 submitted 7 February, 2022; originally announced February 2022.

Comments: The current version of the manuscript has been accepted to IEEE Journal on Selected Areas in Information Theory

Journal ref: IEEE J. Sel. Area. Inf. Theory 3:3 (2022) 481-492

AAPM DL-Sparse-View CT Challenge Submission Report: Designing an Iterative Network for Fanbeam-CT with Unknown Geometry

Authors: Martin Genzel, Jan Macdonald, Maximilian März

Abstract: This report is dedicated to a short motivation and description of our contribution to the AAPM DL-Sparse-View CT Challenge (team name: "robust-and-stable"). The task is to recover breast model phantom images from limited view fanbeam measurements using data-driven reconstruction techniques. The challenge is distinctive in the sense that participants are provided with a collection of ground truth i… ▽ More This report is dedicated to a short motivation and description of our contribution to the AAPM DL-Sparse-View CT Challenge (team name: "robust-and-stable"). The task is to recover breast model phantom images from limited view fanbeam measurements using data-driven reconstruction techniques. The challenge is distinctive in the sense that participants are provided with a collection of ground truth images and their noiseless, subsampled sinograms (as well as the associated limited view filtered backprojection images), but not with the actual forward model. Therefore, our approach first estimates the fanbeam geometry in a data-driven geometric calibration step. In a subsequent two-step procedure, we design an iterative end-to-end network that enables the computation of near-exact solutions. △ Less

Submitted 1 June, 2021; originally announced June 2021.

Comments: This is a technical report of a method participating in a not yet finished challenge. Therefore, it does not contain any final results. In particular, the reported reconstruction errors are only with respect to our own validation split of the provided training data. Once the official challenge report is released, these values will be updated with the results from the actual test set