Understanding Untrained Deep Models for Inverse Problems [From the Guest Editors]: Algorithms and theory
IEEE Signal Processing Magazine · 2026-03 · 43 (2) · 77-92
DOI: 10.1109/msp.2025.3632786 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Magazine, 2026-03. Verfasst von Ismail Alkhouri und 5 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 45 Literaturangaben.
Abstract (Originalsprache)
In recent years, deep learning (DL) methods have been extensively developed for inverse imaging problems (IIPs), encompassing supervised, self-supervised, and generative approaches. Most of these methods require large numbers of labeled or unlabeled training data to learn effective models. However, in many practical applications, such as medical image reconstruction, extensive training datasets are often unavailable or limited. A significant milestone in addressing this challenge came in 2018 with the work of Ulyanov et al. [1], which introduced the deep image prior (DIP)—the first training-data-free convolutional neural network (CNN) method for IIPs. Unlike conventional DL approaches, DIP requires only a CNN, the noisy measurements, and a forward operator. By leveraging the implicit regularization of deep networks initialized with random noise, DIP can learn and restore image structures without relying on external datasets. However, a well-known limitation of DIP is its susceptibility to overfitting, primarily due to overparameterization of the network. In this tutorial article, we provide a comprehensive review of DIP, including a theoretical analysis of its training dynamics. We also categorize and discuss recent advancements in DIP-based methods that are aimed at mitigating overfitting, including techniques such as regularization, network reparameterization, and early stopping. Furthermore, we discuss approaches that combine DIP with pretrained neural networks (NNs), present empirical comparison results against data-centric methods, and highlight open research questions and future directions.
Zitieren
APA
Alkhouri, I., Bell, E., Ghosh, A., Liang, S., Wang, R., & Ravishankar, S. (2026). Understanding Untrained Deep Models for Inverse Problems [From the Guest Editors]: Algorithms and theory. IEEE Signal Processing Magazine, 43(2), 77-92. https://doi.org/10.1109/msp.2025.3632786
Vancouver
Alkhouri I, Bell E, Ghosh A, Liang S, Wang R, Ravishankar S. Understanding Untrained Deep Models for Inverse Problems [From the Guest Editors]: Algorithms and theory. IEEE Signal Processing Magazine. 2026;43(2):77-92. doi:10.1109/msp.2025.3632786
BibTeX
@article{Alkhouri2026,
author = {Alkhouri, Ismail and Bell, Evan and Ghosh, Avrajit and Liang, Shijun and Wang, Rongrong and Ravishankar, Saiprasad},
title = {Understanding Untrained Deep Models for Inverse Problems [From the Guest Editors]: Algorithms and theory},
journal = {IEEE Signal Processing Magazine},
year = {2026},
volume = {43},
number = {2},
pages = {77--92},
doi = {10.1109/msp.2025.3632786}
}
EUVIMED