Wasserstein-based Projections with Applications to Inverse Problems
This provides a theoretically grounded alternative to PnP methods for researchers and practitioners in signal processing and inverse problems, though it is incremental as it builds on existing PnP frameworks.
The paper tackles the limited theoretical guarantees in Plug-and-Play (PnP) methods for inverse problems by introducing Wasserstein-based projections (WPs) that approximate the true projection onto the data manifold with high probability, enabling state-of-the-art unsupervised signal recovery.
Inverse problems consist of recovering a signal from a collection of noisy measurements. These are typically cast as optimization problems, with classic approaches using a data fidelity term and an analytic regularizer that stabilizes recovery. Recent Plug-and-Play (PnP) works propose replacing the operator for analytic regularization in optimization methods by a data-driven denoiser. These schemes obtain state of the art results, but at the cost of limited theoretical guarantees. To bridge this gap, we present a new algorithm that takes samples from the manifold of true data as input and outputs an approximation of the projection operator onto this manifold. Under standard assumptions, we prove this algorithm generates a learned operator, called Wasserstein-based projection (WP), that approximates the true projection with high probability. Thus, WPs can be inserted into optimization methods in the same manner as PnP, but now with theoretical guarantees. Provided numerical examples show WPs obtain state of the art results for unsupervised PnP signal recovery.