DDT: 双分支变形 Transformer 图像去噪（英文版）

英文标题：DDT: Dual-branch Deformable Transformer for Image Denoising中文摘要：本文提出了一种高效的 Dual-branch Deformable Transformer (DT) 去噪网络，在局部和全局分支中均应用了可变形关注机制，能够更好地捕捉图像中的局部和全局交互，并能显著减少计算成本，取得了在真实世界和合成去噪任务方面的最先进性能。英文摘要：Transformer is beneficial for image denoising tasks since it can modellong-range dependencies to overcome the limitations presented by inductiveconvolutional biases. However, directly applying the transformer structure toremove noise is challenging because its complexity grows quadratically wit