发表机构
Fudan University; China Telecom; Harbin Institute of Technology, Weihai（复旦大学; 中国电信; 哈尔滨工业大学（威海））
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出物理引导的光谱蒸馏方法，在资源受限设备上实现高效水下图像增强，保持高性能并提升下游感知任务效果。
AI 中文摘要
水下图像增强对于改善海洋应用中的视觉感知至关重要。现有水下图像增强研究主要关注增强质量和视觉保真度，而很少考虑实时部署能力，而这对于资源受限的水下机器人至关重要。为此，我们提出了一种物理引导的光谱蒸馏（PSD）方法，该方法在保持水下图像增强模型高性能的同时，降低模型容量以实现实时应用。为了分解教师和学生模型的输出，PSD采用多级Haar离散小波变换。它通过频带特定目标传递低频颜色和光照信息以及高频结构细节。此外，PSD的蒸馏过程是退化感知的。我们通过物理头估计退化感知权重，并将其与真实标签引导的可靠性掩码相结合，以选择性地保留有价值的教师指导。在UIEB、LSUI和EUVP数据集上的实验验证了所提方法的有效性。此外，我们展示了增强图像对下游感知任务（包括目标检测）的益处。在自研ROV上的部署进一步证明了其在实际水下场景中的实用适用性。
英文摘要
Underwater image enhancement is crucial for improving visual perception in marine applications. Existing underwater image enhancement studies mainly focus on enhancement quality and visual fidelity, while rarely considering real-time deployment capability, which is essential for resource-constrained underwater robots. To this end, we introduce a physics-guided spectral distillation (PSD) method, which reduces model capacity for real-time applications while maintaining the high performance of underwater image enhancement models. To decompose the outputs of teacher and student models, PSD adopts a multilevel Haar discrete wavelet transform. It transfers low-frequency color and illumination information as well as high-frequency structural details through band-specific objectives. Moreover, the distillation process of PSD is degradation-aware. We estimate degradation-aware weights through a physical head and combine them with ground-truth-guided reliability masks to selectively retain valuable teacher guidance. Experiments on the UIEB, LSUI, and EUVP datasets validate the effectiveness of the proposed method. Furthermore, we demonstrate the benefits of enhanced images for downstream perception tasks, including object detection. Deployment on a self-developed ROV further demonstrates its practical applicability in real-world underwater scenarios.
Comments10 pages, 9 figures
