Infrared and visible image fusion with multi-resolution enhancement and dynamic weighted

Infrared and visible image fusion (IVIF) aims to generate fused images by combining intensity and detailed information from source images. Existing IVIF methods typically rely on manually defined loss function weights and lack dynamic adaptability to complex scenes. Additionally, single-resolution approaches are limited in capturing complementary information across multiple resolutions. In this study, we propose an adaptive loss-weights and multi-resolution enhancement framework for IVIF, improve adaptability to complex scenes by integrating fine-grained details and background textures. Our approach introduces a multi-resolution detail enhancement (MRDE) module that processes extracted features at three distinct scales to enhance texture details. We also present a holographic fusion strategy composed of two modules: the Structural Integration Module (SIM) and Multi-resolution Texture Fusion Modul (MTFM), which effectively integrate structural and texture information across resolutions. Moreover, we propose an adaptive weighting dynamic balancing loss function that adjusts the emphasis on texture and structure during training. Extensive experiments demonstrate that our method outperforms state-of-the-art IVIF techniques, achieving superior visual balance and improved quantitative performance.