Wide Activation for Efficient and Accurate Image Super-Resolution
Wide Activation for Efficient and Accurate Image Super-Resolution
复制标题
DOI：
--
复制
发表时间：
2018-08
期刊：
影响因子：
--
通讯作者：
Jiahui Yu;Yuchen Fan;Jianchao Yang;N. Xu;Zhaowen Wang;Xinchao Wang;Thomas S. Huang
中科院分区：
文献类型：
--
作者：
Jiahui Yu;Yuchen Fan;Jianchao Yang;N. Xu;Zhaowen Wang;Xinchao Wang;Thomas S. Huang
在本报告中，我们证明了在相同的参数和计算预算下，在ReLU激活之前具有更广泛特征的模型在单图像超分辨率（SISR）方面具有更好的性能。由此产生的SR残差网络具有细长的身份映射路径，在每个残差块中激活之前具有更宽的（\（2\times\）至\（4\times\））通道。为了在没有计算开销的情况下进一步扩大激活（\（6\times\）到\（9\times\）），我们将线性低秩卷积引入SR网络，并实现更好的精度-效率权衡。此外，与批量归一化或无归一化相比，我们发现使用权重归一化的训练可以提高深度超分辨率网络的准确性。我们提出的SR网络\textit{WDSR}在大规模DIV 2K图像超分辨率基准测试中取得了更好的结果，在PSNR方面具有相同或更低的计算复杂度。基于WDSR，我们的方法还在所有三个逼真的轨道上获得了NTIRE 2018年单图像超分辨率挑战赛的第一名。实验和消融研究支持广泛激活对图像超分辨率的重要性。代码发布于：此https URL
In this report we demonstrate that with same parameters and computational budgets, models with wider features before ReLU activation have significantly better performance for single image super-resolution (SISR). The resulted SR residual network has a slim identity mapping pathway with wider (\(2\times\) to \(4\times\)) channels before activation in each residual block. To further widen activation (\(6\times\) to \(9\times\)) without computational overhead, we introduce linear low-rank convolution into SR networks and achieve even better accuracy-efficiency tradeoffs. In addition, compared with batch normalization or no normalization, we find training with weight normalization leads to better accuracy for deep super-resolution networks. Our proposed SR network \textit{WDSR} achieves better results on large-scale DIV2K image super-resolution benchmark in terms of PSNR with same or lower computational complexity. Based on WDSR, our method also won 1st places in NTIRE 2018 Challenge on Single Image Super-Resolution in all three realistic tracks. Experiments and ablation studies support the importance of wide activation for image super-resolution. Code is released at: this https URL