007 论文：Restormer: Efficient Transformer for High-Resolution Image Restoration

提要：主要是针对课程作业的要求，选了一篇cvpr 2022的论文+代码进行阅读；

预期完成时间：24.10.31-24.11.03

Abstract

Since convolutional neural networks (CNNs) perform well at learning generalizable image
priors from largescale data, these models have been extensively applied to image restoration
and related tasks.
从大规模数据中学习图像先验：通过对大量图像数据的训练，使得卷积神经网络（CNN）能够提取和掌握一些图像的基本特征和规律，这些特征在不同的图像中可能会重复出现。

Recently, another class of neural architectures, Transformers, have shown
significant performance gains on natural language and high-level vision tasks. While the
Transformer model mitigates the shortcomings of CNNs (i.e., limited receptive field and
inadaptability to input content), its computational complexity grows quadratically with the
spatial resolution, therefore making it infeasible to apply to most image restoration tasks
involving high-resolution images.
Transformer架构在自然语言处理和高层次视觉任务中的显著性能提升。虽然Transformer模型克服了卷积神经网络（CNN）的局限性（如感受野有限和对输入内容的适应性差），但其计算复杂度随着空间分辨率的增加而呈二次增长。这使得在涉及高分辨率图像的图像修复任务中应用Transformer变得不可行。

In this work, we propose an efficient Transformer model by making several key designs in the
building blocks (multi-head attention and feed-forward network) such that it can capture long-range pixel interactions, while still remaining applicable to large images. Our model, named Restoration Transformer (Restormer), achieves state-of-the-art results on several image restoration tasks, including image deraining, single-image motion deblurring, defocus deblurring (single-image and dual-pixel data), and image denoising (Gaussian grayscale/color denoising, and real image denoising).
Restoration Transformer（Restormer）是一种高效的Transformer模型，通过优化多头注意力和前馈网络设计，实现了长距离像素交互的捕捉，适用于大图像处理。该模型在多个图像恢复任务中（如图像去雨、运动去模糊、焦点去模糊和图像去噪）表现出色，取得了最先进的结果。