{"ID":2876404,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.00346","arxiv_id":"2509.00346","title":"LUT-Fuse: Towards Extremely Fast Infrared and Visible Image Fusion via Distillation to Learnable Look-Up Tables","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.","short_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 designe...","url_abs":"https://arxiv.org/abs/2509.00346","url_pdf":"https://arxiv.org/pdf/2509.00346v1","authors":"[\"Xunpeng Yi\",\"Yibing Zhang\",\"Xinyu Xiang\",\"Qinglong Yan\",\"Han Xu\",\"Jiayi Ma\"]","published":"2025-08-30T03:59:26Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false,"code_links":[{"ID":610291,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2876404,"paper_url":"https://arxiv.org/abs/2509.00346","paper_title":"LUT-Fuse: Towards Extremely Fast Infrared and Visible Image Fusion via Distillation to Learnable Look-Up Tables","repo_url":"https://github.com/zyb5/LUT-Fuse","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}