MobileFusion: Mobile-Friendly Infrared and Visible Image Fusion via Structural Re-parameterization

Deep neural networks have recently advanced infrared and visible image fusion (IVIF), but most existing methods rely on sophisticated yet redundant designs, which hinder real-time deployment on mobile devices with limited compute and memory. In this paper, we present MobileFusion, an extremely lightweight and effective convolutional framework that achieves high-quality fusion under strict resource constraints. MobileFusion leverages a novel re-parameterizable multi-branch convolution module to promote cross-modal interactions during training while collapsing into a single-path operator for fast inference. It further incorporates a lightweight attention module to enhance context awareness, together with a re-parameterized feed-forward network to improve feature expressiveness. Extensive experiments demonstrate that MobileFusion delivers a favorable trade-off between fusion quality and computational efficiency, enabling real-time and high-quality IVIF on resource-constrained platforms. The source code is available at https://github.com/sucessfullys/MobileFusion.

Infrared cameras can reveal warm objects such as people or vehicles in darkness, fog, or low-light scenes, while visible cameras capture rich textures and fine visual details. Combining these two types of images can help machines and humans better understand complex environments. However, many existing image fusion methods are too large and slow to run efficiently on mobile or edge devices, which limits their use in real-time applications such as autonomous driving, surveillance, and portable sensing systems. We propose MobileFusion, a very lightweight method for combining infrared and visible images. During training, the model uses multiple branches to learn how to mix useful information from both image types. During deployment, these branches are merged into a single efficient operation, so the model can run much faster while keeping strong fusion quality. We also add lightweight attention and feature enhancement modules to better preserve important targets, edges, textures, and structural details. Our results show that MobileFusion achieves high-quality image fusion with very few parameters and low latency. This makes real-time infrared-visible image fusion more practical for resource-limited devices.