训练集
-
91 images:来自 J. Yang, J. Wright, T. S. Huang, and Y. Ma. Image superresolution via sparse representation. TIP, 2010.
-
291 images:来自D. Martin, C. Fowlkes, D. Tal, and J. Malik. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics. In ICCV, 2001.
-
ImageNet:SRCNN中采用该训练集进行训练。
测试集
-
Set5:来自C. G. Marco Bevilacqua, Aline Roumy and M.-L. A. Morel. Low-complexity single-image super-resolution based on nonnegative neighbor embedding. In BMVC, 2012.
-
Set14:来自R. Zeyde, M. Elad, and M. Protter. On single image scale-upusing sparse-representations. In Curves and Surfaces, pages 711–730. Springer, 2012.
-
Urban100:来自J.-B. Huang, A. Singh, and N. Ahuja. Single image superresolution using transformed self-exemplars. In CVPR, 2015.
-
B100:来自R. Timofte, V. De Smet, and L. Van Gool. A+: Adjusted anchored neighborhood regression for fast super-resolution. In ACCV, 2014.
测试指标
- PSNR: Peak Signal to Noise Ratio,峰值信噪比.
- SSIM: structural similarity index,结构相似性.
- MOS: mean-opinion-score,平均主观意见分