Improving Decoupled Posterior Sampling for Inverse Problems using Data Consistency Constraint
This work addresses errors in posterior sampling for inverse problems, which is important for applications like image reconstruction, though it appears incremental as it builds on existing decoupled methods.
The paper tackled errors in diffusion models for solving inverse problems by proposing Guided Decoupled Posterior Sampling (GDPS), which integrates a data consistency constraint to improve convergence, achieving state-of-the-art performance on FFHQ and ImageNet datasets across various tasks.
Diffusion models have shown strong performances in solving inverse problems through posterior sampling while they suffer from errors during earlier steps. To mitigate this issue, several Decoupled Posterior Sampling methods have been recently proposed. However, the reverse process in these methods ignores measurement information, leading to errors that impede effective optimization in subsequent steps. To solve this problem, we propose Guided Decoupled Posterior Sampling (GDPS) by integrating a data consistency constraint in the reverse process. The constraint performs a smoother transition within the optimization process, facilitating a more effective convergence toward the target distribution. Furthermore, we extend our method to latent diffusion models and Tweedie's formula, demonstrating its scalability. We evaluate GDPS on the FFHQ and ImageNet datasets across various linear and nonlinear tasks under both standard and challenging conditions. Experimental results demonstrate that GDPS achieves state-of-the-art performance, improving accuracy over existing methods.