Wide Activation for Efficient Image and Video Super-Resolution

Jiahui Yu (University of Illinois at Urbana-Champaign), Yuchen Fan (University of Illinois at Urbana-Champaign), Thomas Huang (University of Illinois at Urbana-Champaign)

Abstract
In this work we demonstrate that with same parameters and computational budgets, models with wider features before ReLU activation have significantly better performance for image and video super-resolution. 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, under same or lower computational complexity. Based on WDSR, our method won \textbf{1st places} in \textit{NTIRE 2018 Challenge on Single Image Super-Resolution} in all three realistic tracks. Moreover, a simple frame-concatenation based WDSR achieved \textbf{2nd places} in three out of four tracks of \textit{NTIRE 2019 Challenge for Video Super-Resolution and Deblurring}. Our experiments and ablation studies support the importance of wide activation. Code and models will be publicly available.

DOI
10.5244/C.33.52
https://dx.doi.org/10.5244/C.33.52

Files
Paper (PDF)

BibTeX