Reversible Efficient Diffusion for Image Fusion

2026-01-28

AI Summary

This work addresses the limitations of conventional diffusion models in multimodal image fusion, where the Markovian process accumulates noise errors, leading to detail loss and inefficient end-to-end explicit supervised training. To overcome these issues, the paper proposes a Reversible Efficient Diffusion model (RED) that leverages a reversible architecture to enable efficient explicit supervision without explicitly modeling the data distribution. By mitigating noise accumulation, RED significantly enhances detail preservation and visual fidelity in fused images while maintaining high generation quality and computational efficiency.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

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.

Problem

• image fusion
• diffusion models
• detail loss
• computational efficiency
• multi-modal

Innovation

• Reversible Efficient Diffusion
• image fusion
• diffusion models
• explicit supervision
• noise error reduction