TL;DR本文中，我们提出了一种基于CNN的Sparse Mask SR (SMSR) 网络，以学习稀疏掩模，以便在保持可比较性能的同时，准确地定位和跳过冗余计算，从而提高SR网络的推理效率。
Abstract
Current cnn-based super-resolution (SR) methods process all locations equally with computational resources being uniformly assigned in space. However, since highfrequency details mainly lie around edges and textu