BriefGPT.xyz
Nov, 2015
使用深度卷积网络进行精确图像超分辨率
Accurate Image Super-Resolution Using Very Deep Convolutional Networks
HTML
PDF
Jiwon Kim, Jung Kwon Lee, Kyoung Mu Lee
TL;DR
本研究提出了一种用于单幅图像超分辨率的高精度方法,利用 VGG-net 进行训练,使用20个卷积层避免过拟合,训练方法基于小滤波器,只学习残差,并使用极高的学习率。实验表明,该方法在精度上优于现有方法。
Abstract
We present a highly accurate
single-image super-resolution
(SR) method. Our method uses a very deep
convolutional network
inspired by
vgg-net
→