Reversible Efficient Diffusion for Image Fusion
This work addresses a specific bottleneck in image fusion for applications requiring detailed and consistent fused images, representing an incremental improvement over existing diffusion-based methods.
The paper tackles the problem of detail loss in diffusion models for multi-modal image fusion by proposing the Reversible Efficient Diffusion (RED) model, which uses explicit supervision to improve computational efficiency and maintain high visual fidelity.
Multi-modal image fusion aims to consolidate complementary information from diverse source images into a unified representation. The fused image is expected to preserve fine details and maintain high visual fidelity. While diffusion models have demonstrated impressive generative capabilities in image generation, they often suffer from detail loss when applied to image fusion tasks. This issue arises from the accumulation of noise errors inherent in the Markov process, leading to inconsistency and degradation in the fused results. However, incorporating explicit supervision into end-to-end training of diffusion-based image fusion introduces challenges related to computational efficiency. To address these limitations, we propose the Reversible Efficient Diffusion (RED) model - an explicitly supervised training framework that inherits the powerful generative capability of diffusion models while avoiding the distribution estimation.