不用卷积的图像去噪模型，靠池化捕捉空间信息，参数更少、效果更好。
Pooling-Based Context Modeling for Convolution-Free Deep Image Prior
- 用池化替代卷积，高效建模图像局部上下文
- 在多个数据集上达到可比去噪效果，参数量大幅降低
- 适合追求轻量化与高稳定性的图像修复任务
卷积神经网络（CNN）通过利用邻近像素的空间上下文实现强大的去噪性能。深度图像先验（DIP）借助这一特性，仅凭单张含噪图像即可恢复图像，无需大规模训练数据。然而，DIP 的超参数化架构常导致优化过程中拟合噪声。本文提出 Pool-DIP，一种无卷积结构，通过基于池化的对比建模有效捕捉空间上下文。相比传统卷积型 DIP 模型，Pool-DIP 显著减少参数量与计算复杂度，同时提升去噪性能。实验表明，其在多个数据集（包括真实世界基准）上表现优异。谱分析进一步揭示，Pool-DIP 在优化过程中稳定了高频分量的演化，并抑制错误高频信号。该架构还可泛化至超分辨率与修补等其他图像修复任务。
原文摘要 · Abstract (English)
Convolutional Neural Networks (CNNs) achieve strong denoising performance by exploiting spatial context from neighboring pixels. Deep Image Prior (DIP) leverages this property to restore images from a single noisy input without requiring large datasets. However, the over-parameterized architecture of DIP often leads to noise fitting during optimization. In this paper, we propose Pool-DIP, a convolution-free architecture that incorporates pooling-based contrast modeling to capture spatial context efficiently. Pool-DIP improves denoising performance while significantly reducing the number of parameters and computational complexity compared to convolution-based DIP models. Experimental results show that Pool-DIP achieves competitive performance across multiple datasets, including a real-world benchmark. Spectral analysis further reveals that Pool-DIP stabilizes the evolution of high-frequency components during optimization and suppresses erroneous high-frequency signals. The proposed architecture also generalizes well to other image restoration tasks such as super-resolution and inpainting.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。