# LUT-Fuse: Towards Extremely Fast Infrared and Visible Image Fusion via Distillation to Learnable Look-Up Tables

## TL;DR

LUT-Fuse introduces a novel, extremely fast infrared and visible image fusion method using learnable lookup tables and distillation, achieving high speed and performance suitable for real-time applications on low-power devices.

## Contribution

The paper presents a new LUT-based fusion model with a distillation strategy, significantly improving speed and efficiency over existing methods for multi-modal image fusion.

## Key findings

- Requires less than one-tenth of the time of state-of-the-art algorithms
- Achieves high operational speed on low-power devices
- Demonstrates superior fusion performance and stability

## Abstract

Current advanced research on infrared and visible image fusion primarily focuses on improving fusion performance, often neglecting the applicability on real-time fusion devices. In this paper, we propose a novel approach that towards extremely fast fusion via distillation to learnable lookup tables specifically designed for image fusion, termed as LUT-Fuse. Firstly, we develop a look-up table structure that utilizing low-order approximation encoding and high-level joint contextual scene encoding, which is well-suited for multi-modal fusion. Moreover, given the lack of ground truth in multi-modal image fusion, we naturally proposed the efficient LUT distillation strategy instead of traditional quantization LUT methods. By integrating the performance of the multi-modal fusion network (MM-Net) into the MM-LUT model, our method achieves significant breakthroughs in efficiency and performance. It typically requires less than one-tenth of the time compared to the current lightweight SOTA fusion algorithms, ensuring high operational speed across various scenarios, even in low-power mobile devices. Extensive experiments validate the superiority, reliability, and stability of our fusion approach. The code is available at https://github.com/zyb5/LUT-Fuse.