Weighted total variation regularization for inverse problems with significant null spaces

TL;DR

This paper extends weighted total variation regularization to inverse problems with large null spaces, improving the recovery of spatially extended sources away from data acquisition sites, especially in EEG and ECG applications.

Contribution

The paper introduces a weighted TV regularization method for inverse problems with large null spaces, supported by analysis and numerical experiments, and explores a hybrid approach combining sparsity and TV.

Findings

Weighted TV regularization successfully recovers large, piecewise constant sources away from boundaries.

The hybrid weighted-sparsity and TV approach captures both small and large sources.

Weighted TV improves source localization in inverse problems with significant null spaces.

Abstract

We consider inverse problems with large null spaces, which arise in important applications such as in inverse ECG and EEG procedures. Standard regularization methods typically produce solutions in or near the orthogonal complement of the forward operator's null space. This often leads to inadequate results, where internal sources are mistakenly interpreted as being near the data acquisition sites -- e.g., near or at the body surface in connection with EEG and ECG recordings. To mitigate this, we previously proposed weighting schemes for Tikhonov and sparsity regularization. Here, we extend this approach to total variation (TV) regularization, which is particularly suited for identifying spatially extended regions with approximately constant values. We introduce a weighted TV-regularization method, provide supporting analysis, and demonstrate its performance through numerical experiments. Unlike standard TV regularization, the weighted version successfully recovers the location and size of large, piecewise constant sources away from the boundary, though not their exact shape. Additionally, we explore a hybrid weighted-sparsity and TV regularization approach, which better captures both small and large sources, albeit with somewhat more blurred reconstructions than the weighted TV method alone.