Domain Expansion via Network Adaptation for Solving Inverse Problems

TL;DR

This paper introduces a flexible domain adaptation framework that enhances pretrained networks for inverse problems, improving robustness and performance across various imaging tasks under domain shifts.

Contribution

The paper proposes a novel, parameter-efficient domain adaptation method for inverse problems, addressing robustness issues caused by data, measurement, and noise shifts.

Findings

Significantly improved performance over existing methods.

Enhanced robustness to domain, measurement, and noise shifts.

Parameter efficiency in adaptation process.

Abstract

Deep learning-based methods deliver state-of-the-art performance for solving inverse problems that arise in computational imaging. These methods can be broadly divided into two groups: (1) learn a network to map measurements to the signal estimate, which is known to be fragile; (2) learn a prior for the signal to use in an optimization-based recovery. Despite the impressive results from the latter approach, many of these methods also lack robustness to shifts in data distribution, measurements, and noise levels. Such domain shifts result in a performance gap and in some cases introduce undesired artifacts in the estimated signal. In this paper, we explore the qualitative and quantitative effects of various domain shifts and propose a flexible and parameter efficient framework that adapt pretrained networks to such shifts. We demonstrate the effectiveness of our method for a number of natural image, MRI, and CT reconstructions tasks under domain, measurement model, and noise-level shifts. Our experiments demonstrate that our method provides significantly better performance and parameter efficiency compared to existing domain adaptation techniques.