Open Access

ARTICLE

An Infrared-Visible Image Fusion Network with Channel-Switching for Low-Light Object Detection

Tianzhe Jiao, Yuming Chen, Xiaoyue Feng, Chaopeng Guo, Jie Song^*

Software College, Northeastern University, Shenyang, 110819, China

* Corresponding Author: Jie Song. Email:

Computers, Materials & Continua 2025, 85(2), 2681-2700. https://doi.org/10.32604/cmc.2025.069235

Received 18 June 2025; Accepted 15 August 2025; Issue published 23 September 2025

Abstract

Visible-infrared object detection leverages the day-night stable object perception capability of infrared images to enhance detection robustness in low-light environments by fusing the complementary information of visible and infrared images. However, the inherent differences in the imaging mechanisms of visible and infrared modalities make effective cross-modal fusion challenging. Furthermore, constrained by the physical characteristics of sensors and thermal diffusion effects, infrared images generally suffer from blurred object contours and missing details, making it difficult to extract object features effectively. To address these issues, we propose an infrared-visible image fusion network that realizes multimodal information fusion of infrared and visible images through a carefully designed multi-scale fusion strategy. First, we design an adaptive gray-radiance enhancement (AGRE) module to strengthen the detail representation in infrared images, improving their usability in complex lighting scenarios. Next, we introduce a channel-spatial feature interaction (CSFI) module, which achieves efficient complementarity between the RGB and infrared (IR) modalities via dynamic channel switching and a spatial attention mechanism. Finally, we propose a multi-scale enhanced cross-attention fusion (MSECA) module, which optimizes the fusion of multi-level features through dynamic convolution and gating mechanisms and captures long-range complementary relationships of cross-modal features on a global scale, thereby enhancing the expressiveness of the fused features. Experiments on the KAIST, M3FD, and FLIR datasets demonstrate that our method delivers outstanding performance in daytime and nighttime scenarios. On the KAIST dataset, the miss rate drops to 5.99%, and further to 4.26% in night scenes. On the FLIR and M3FD datasets, it achieves scores of 79.4% and 88.9%, respectively.

---

Keywords

---