ShiftLUT：面向高效图像复原的空间偏移增强查找表
ShiftLUT: Spatial Shift Enhanced Look-Up Tables for Efficient Image Restoration
用可学习通道级空间偏移（LSS）在不增加 LUT 体积的前提下扩展感受野，配合非对称双分支把算力倾斜到信息密集分支以降低推理延迟，这套思路可直接迁移到任何基于 LUT 或轻量卷积的端侧图像复原/增强模型设计。
原文依据 Learnable Spatial Shift module (LSS) is introduced to expand the receptive field by applying learnable, channel-wise spatial offsets on feature maps.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论基于查找表的方法已成为高效图像复原任务中一个有前景的方向。近期基于查找表的方法主要通过扩大感受野来提升性能，但这不可避免地引入额外的计算与存储开销，阻碍了其在边缘设备上的部署。为解决该问题，本文提出 ShiftLUT，一种在所有基于查找表的方法中取得最大感受野同时保持高效率的新框架。我们的关键洞见在于三个互补的组件。首先，引入可学习空间偏移模块（LSS），通过在特征图上施加可学习的、逐通道的空间偏移来扩大感受野。其次，我们提出一种非对称双分支架构，将更多计算分配给信息密集的分支，在不损害复原质量的前提下大幅降低推理延迟。最后，我们引入一种特征级查找表压缩策略，称为误差有界自适应采样（EAS），以最小化存储开销。与先前的最优方法 TinyLUT 相比，ShiftLUT 在多个标准基准上取得 3.8 倍更大的感受野，平均峰值信噪比提升超过 0.21 dB，同时保持较小的存储规模和推理时间。
英文原文摘要 1297 字符
Look-Up Table based methods have emerged as a promising direction for efficient image restoration tasks. Recent LUT-based methods focus on improving their performance by expanding the receptive field. However, they inevitably introduce extra computational and storage overhead, which hinders their deployment in edge devices.To address this issue, we propose ShiftLUT, a novel framework that attains the largest receptive field among all LUT-based methods while maintaining high efficiency. Our key insight lies in three complementary components. First, Learnable Spatial Shift module (LSS) is introduced to expand the receptive field by applying learnable, channel-wise spatial offsets on feature maps. Second, we propose an asymmetric dual-branch architecture that allocates more computation to the information-dense branch, substantially reducing inference latency without compromising restoration quality.Finally, we incorporate a feature-level LUT compression strategy called Error-bounded Adaptive Sampling (EAS) to minimize the storage overhead.Compared to the previous state-of-the-art method TinyLUT, ShiftLUT achieves a 3.8xlarger receptive field and improves an average PSNR by over 0.21 dB across multiple standard benchmarks, while maintaining a small storage size and inference time.