Hybrid Fusion: One-Minute Efficient Training for Zero-Shot Cross-Domain Image Fusion

TL;DR

This paper introduces a hybrid image fusion method combining deep learning and classical techniques, enabling fast, high-quality, zero-shot full-resolution fusion across diverse domains with minimal training time.

Contribution

A novel hybrid framework that decouples policy learning from pixel synthesis, achieving efficient full-resolution training and state-of-the-art performance in zero-shot image fusion.

Findings

Achieves SOTA-like performance in about one minute on high-end GPU

Demonstrates powerful zero-shot generalization across diverse tasks

Eliminates the train-inference gap in image fusion models

Abstract

Image fusion seeks to integrate complementary information from multiple sources into a single, superior image. While traditional methods are fast, they lack adaptability and performance. Conversely, deep learning approaches achieve state-of-the-art (SOTA) results but suffer from critical inefficiencies: their reliance on slow, resource-intensive, patch-based training introduces a significant gap with full-resolution inference. We propose a novel hybrid framework that resolves this trade-off. Our method utilizes a learnable U-Net to generate a dynamic guidance map that directs a classic, fixed Laplacian pyramid fusion kernel. This decoupling of policy learning from pixel synthesis enables remarkably efficient full-resolution training, eliminating the train-inference gap. Consequently, our model achieves SOTA-comparable performance in about one minute on a RTX 4090 or two minutes on a consumer laptop GPU from scratch without any external model and demonstrates powerful zero-shot generalization across diverse tasks, from infrared-visible to medical imaging. By design, the fused output is linearly constructed solely from source information, ensuring high faithfulness for critical applications. The codes are available at https://github.com/Zirconium233/HybridFusion