Measurement-Aligned Sampling for Inverse Problem
This addresses a key limitation in inverse problem solving for applications like imaging or signal processing, though it appears incremental as it builds on existing approaches like DDNM and TMPD.
The paper tackles the challenge of incorporating conflicting prior and measurement signals in diffusion models for inverse problems, especially under non-Gaussian or unknown noise, by proposing Measurement-Aligned Sampling (MAS), which outperforms state-of-the-art methods across various tasks.
Diffusion models provide a powerful way to incorporate complex prior information for solving inverse problems. However, existing methods struggle to correctly incorporate guidance from conflicting signals in the prior and measurement, and often failed to maximizing the consistency to the measurement, especially in the challenging setting of non-Gaussian or unknown noise. To address these issues, we propose Measurement-Aligned Sampling (MAS), a novel framework for linear inverse problem solving that flexibly balances prior and measurement information. MAS unifies and extends existing approaches such as DDNM, TMPD, while generalizing to handle both known Gaussian noise and unknown or non-Gaussian noise types. Extensive experiments demonstrate that MAS consistently outperforms state-of-the-art methods across a variety of tasks, while maintaining relatively low computational cost.