Direct Diffusion Bridge using Data Consistency for Inverse Problems
This work addresses the problem of slow inference in diffusion-based inverse problem solvers for researchers and practitioners in computational imaging or signal processing, representing an incremental improvement over existing methods.
The paper tackles the speed limitation of diffusion model-based inverse problem solvers by proposing a modified inference procedure called data Consistent Direct Diffusion Bridge (CDDB), which imposes data consistency without fine-tuning and achieves state-of-the-art results in perception and distortion metrics.
Diffusion model-based inverse problem solvers have shown impressive performance, but are limited in speed, mostly as they require reverse diffusion sampling starting from noise. Several recent works have tried to alleviate this problem by building a diffusion process, directly bridging the clean and the corrupted for specific inverse problems. In this paper, we first unify these existing works under the name Direct Diffusion Bridges (DDB), showing that while motivated by different theories, the resulting algorithms only differ in the choice of parameters. Then, we highlight a critical limitation of the current DDB framework, namely that it does not ensure data consistency. To address this problem, we propose a modified inference procedure that imposes data consistency without the need for fine-tuning. We term the resulting method data Consistent DDB (CDDB), which outperforms its inconsistent counterpart in terms of both perception and distortion metrics, thereby effectively pushing the Pareto-frontier toward the optimum. Our proposed method achieves state-of-the-art results on both evaluation criteria, showcasing its superiority over existing methods. Code is available at https://github.com/HJ-harry/CDDB