arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2203.06697cs.CV
收藏

一种有效的图像超分辨率远程注意力网络

Efficient Long-Range Attention Network for Image Super-resolution

The Hong Kong Polytechnic University(香港理工大学);OPPO Research(OPPO研究院)

Xindong Zhang, Hui Zeng, Shi Guo, Lei Zhang

摘要:

最近,基于变换的方法已经在各种视觉任务中表现出令人印象深刻的结果,包括图像超分辨率(SR),通过利用自注意(SA)的特征提取。然而,在大多数现有的基于Transformer的模型中,SA的计算是非常昂贵的,而一些所采用的操作对于SR任务来说可能是冗余的。这限制了SA计算的范围,并因此限制了SR性能。在这项工作中,我们提出了一个有效的远程注意网络(ELAN)的图像SR。具体来说,我们首先采用移位卷积(shift-conv)来有效地提取图像局部结构信息,同时保持与1x1卷积相同的复杂度,然后提出了一种分组多尺度自注意(GMSA)模块,其使用不同的窗口大小来计算关于非重叠的特征组的SA,以利用远程图像依赖性。一个高效的远程注意力块(ELAB),然后建立简单的级联两个shift-conv与GMSA模块,这是进一步加速使用共享的注意力机制。没有花里胡哨,我们的ELAN遵循一个相当简单的设计,通过顺序级联ELAB。大量的实验表明,ELAN获得更好的结果对基于变压器的SR模型,但具有显着降低的复杂性。源代码可以在www.example.com上找到。

英文摘要:

Recently, transformer-based methods have demonstrated impressive results in various vision tasks, including image super-resolution (SR), by exploiting the self-attention (SA) for feature extraction. However, the computation of SA in most existing transformer based models is very expensive, while some employed operations may be redundant for the SR task. This limits the range of SA computation and consequently the SR performance. In this work, we propose an efficient long-range attention network (ELAN) for image SR. Specifically, we first employ shift convolution (shift-conv) to effectively extract the image local structural information while maintaining the same level of complexity as 1x1 convolution, then propose a group-wise multi-scale self-attention (GMSA) module, which calculates SA on non-overlapped groups of features using different window sizes to exploit the long-range image dependency. A highly efficient long-range attention block (ELAB) is then built by simply cascading two shift-conv with a GMSA module, which is further accelerated by using a shared attention mechanism. Without bells and whistles, our ELAN follows a fairly simple design by sequentially cascading the ELABs. Extensive experiments demonstrate that ELAN obtains even better results against the transformer-based SR models but with significantly less complexity. The source code can be found at https://github.com/xindongzhang/ELAN.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑