IAF-Net: Illumination-Adaptive Fusion for Low-Light Urban Road Segmentation
This work provides a solution for improving the robustness of semantic road segmentation for autonomous driving systems operating in low-light and adverse weather conditions, which is an incremental improvement for existing methods.
This paper addresses the problem of semantic road segmentation under low-light conditions, where existing methods struggle due to illumination-dependent changes in modality reliability. The proposed IAF-Net dynamically adjusts fusion weights of RGB and geometric features, achieving state-of-the-art overall performance on the nuScenes Nighttime Road Segmentation dataset and demonstrating robustness across adverse weather conditions on the CARLA Multi-Weather Road Segmentation dataset. The core Illumination-Adaptive Fusion (IAF) module provides a significant gain of 0.70% in MaxF.
Semantic road segmentation is important for autonomous driving, but existing methods suffer severe performance degradation under low-light conditions. Many existing multi-modal fusion methods do not explicitly adapt to illumination-dependent changes in modality reliability, which can propagate degraded RGB features into the fused representation at night. We propose IAF-Net (Illumination-Adaptive Fusion Network), an end-to-end framework with illumination-adaptive fusion for robust road segmentation across different lighting conditions. It dynamically adjusts fusion weights of RGB and geometric features via the core Illumination-Adaptive Fusion (IAF) module, and enhances low-light feature selection with a brightness-modulated attention decoder. We also construct two dedicated datasets: nuScenes Nighttime Road Segmentation (nuScenes-NRS) and CARLA Multi-Weather Road Segmentation (CARLA-MWRS). Experiments on nuScenes-NRS show state-of-the-art overall performance among the compared methods, while CARLA-MWRS further validates robustness across adverse weather conditions. Ablation studies on a 40% training subset further highlight the importance of the IAF module, which provides the largest individual gain of 0.70% in MaxF.