2 January 2023 in IEEE Signal Processing Magazine
In recent years, deep learning has achieved remarkable empirical success for image reconstruction. This has catalyzed an ongoing quest for the precise characterization of the correctness and reliability of data-driven methods in critical use cases, for instance, in medical imaging. Notwithstanding the excellent performance and efficacy of deep learning-based methods, concerns have been raised regarding the approaches’ stability, or lack thereof, with serious practical implications. Significant advances have been made in recent years to unravel the inner workings of data-driven image recovery methods, challenging their widely perceived black-box nature. In this article, we specify relevant notions of convergence for data-driven image reconstruction, which forms the basis of a survey of learned methods with mathematically rigorous reconstruction guarantees. An example that is highlighted is the role of input-convex neural networks (ICNNs), offering the possibility to combine the power of deep learning with classical convex regularization theory for devising methods that are provably convergent. This survey article is aimed at both methodological researchers seeking to advance the frontiers of our understanding of data-driven image reconstruction methods as well as practitioners by providing an accessible description of useful convergence concepts and by placing some of the existing empirical practices on a solid mathematical foundation.
2 April 2024 in SIAM Journal on Imaging Sciences
CSCarola-Bibiane Schönlieb
Plug-and-Play (PnP) methods are a class of efficient iterative methods that aim to combine data fidelity terms and deep denoisers using classical optimization algorithms, such as ISTA or ADMM, with applications in inverse problems and imaging. Provable PnP methods are a subclass of PnP methods with convergence guarantees, such as fixed point convergence or convergence to critical points of some energy function. Many existing provable PnP methods impose heavy restrictions on the denoiser or fidelity function, such as nonexpansiveness or strict convexity, respectively. In this work, we propose a novel algorithmic approach incorporating quasi-Newton steps into a provable PnP framework based on proximal denoisers, resulting in greatly accelerated convergence while retaining light assumptions on the denoiser. By characterizing the denoiser as the proximal operator of a weakly convex function, we show that the fixed points of the proposed quasi-Newton PnP algorithm are critical points of a weakly convex function. Numerical experiments on image deblurring and super-resolution demonstrate 2–8x faster convergence as compared to other provable PnP methods with similar reconstruction quality.
7 February 2024 in Computer Methods and Programs in Biomedicine
CSCarola-Bibiane Schönlieb
4D flow magnetic resonance imaging provides time-resolved blood flow velocity measurements, but suffers from limitations in spatio-temporal resolution and noise. In this study, we investigated the use of sinusoidal representation networks (SIRENs) to improve denoising and super-resolution of velocity fields measured by 4D flow MRI in the thoracic aorta. Efficient training of SIRENs in 4D was achieved by sampling voxel coordinates and enforcing the no-slip condition at the vessel wall. A set of synthetic measurements were generated from computational fluid dynamics simulations, reproducing different noise levels. The influence of SIREN architecture was systematically investigated, and the performance of our method was compared to existing approaches for 4D flow denoising and super-resolution. Compared to existing techniques, a SIREN with 300 neurons per layer and 20 layers achieved lower errors (up to 50% lower vector normalized root mean square error, 42% lower magnitude normalized root mean square error, and 15% lower direction error) in velocity and wall shear stress fields. Applied to real 4D flow velocity measurements in a patient-specific aortic aneurysm, our method produced denoised and super-resolved velocity fields while maintaining accurate macroscopic flow measurements. This study demonstrates the feasibility of using SIRENs for complex blood flow velocity representation from clinical 4D flow, with quick execution and straightforward implementation.
3 June 2024 in Foundations of Computational Mathematics
Regularization is necessary when solving inverse problems to ensure the well-posedness of the solution map. Additionally, it is desired that the chosen regularization strategy is convergent in the sense that the solution map converges to a solution of the noise-free operator equation. This provides an important guarantee that stable solutions can be computed for all noise levels and that solutions satisfy the operator equation in the limit of vanishing noise. In recent years, reconstructions in inverse problems are increasingly approached from a data-driven perspective. Despite empirical success, the majority of data-driven approaches do not provide a convergent regularization strategy. One such popular example is given by iterative plug-and-play (PnP) denoising using off-the-shelf image denoisers. These usually provide only convergence of the PnP iterates to a fixed point, under suitable regularity assumptions on the denoiser, rather than convergence of the method as a regularization technique, thatis under vanishing noise and regularization strength. This paper serves two purposes: first, we provide an overview of the classical regularization theory in inverse problems and survey a few notable recent data-driven methods that are provably convergent regularization schemes. We then continue to discuss PnP algorithms and their established convergence guarantees. Subsequently, we consider PnP algorithms with learned linear denoisers and propose a novel spectral filtering technique of the denoiser to control the strength of regularization. Further, by relating the implicit regularization of the denoiser to an explicit regularization functional, we are the first to rigorously show that PnP with a learned linear denoiser leads to a convergent regularization scheme. The theoretical analysis is corroborated by numerical experiments for the classical inverse problem of tomographic image reconstruction.
17 January 2022 in Journal of the Science of Food and Agriculture
8 January 2018 in IEEE Signal Processing Letters
We consider the problem of signal reconstruction from quadratic measurements that are encoded as +1 or -1 depending on whether they exceed a predetermined positive threshold or not. Binary measurements are fast to acquire and inexpensive in terms of hardware. We formulate the problem of signal reconstruction using a consistency criterion, wherein one seeks to find a signal that is in agreement with the measurements. To enforce consistency, we construct a convex cost using a one-sided quadratic penalty and minimize it using an iterative accelerated projected gradient-descent (APGD) technique. The PGD scheme reduces the cost function in each iteration, whereas incorporating momentum into PGD, notwithstanding the lack of such a descent property, exhibits faster convergence than PGD empirically. We refer to the resulting algorithm as binary phase retrieval (BPR). Considering additive white noise contamination prior to quantization, we also derive the Cramer-Rao Bound (CRB) for the binary encoding model. Experimental results demonstrate that the BPR algorithm yields a signal-to-reconstruction error ratio (SRER) of approximately 25 dB in the absence of noise. In the presence of noise prior to quantization, the SRER is within 2 to 3 dB of the CRB.