search

greatlog / DAN

This is an official implementation of Unfolding the Alternating Optimization for Blind Super Resolution
231 stars 41 forks source link

请问测试集是怎么生成的 #21

Closed Fangzhenxuan closed 3 years ago

Fangzhenxuan commented 3 years ago

您好，对于X2的SR，请问怎么生成Gaussian8的kernel，Set5的每张测试图片都需要生成8张不同的LR图片吗

greatlog commented 3 years ago

是的，对Setting1，每张图都需要生成8张不同的LR，分别对应8个不同Gaussian Kernel. 对x4来说，kerne的标准差在[0.2, 4.0]等间隔取8个，x3在[0.2, 3.0]等间隔取8个, x2在[0.2, 2.0]等间隔取8个。

Fangzhenxuan commented 3 years ago

非常感谢您的回复，但是您说的范围是training的吧。对于testing，x4来说，kerne的标准差应该在[1.8, 3.2]等间隔取8个，x3在[1.35, 2.40]等间隔取8个, x2在[0.80, 1.60]等间隔取8个， ‘’ we uniformly choose 8 kernels, denoted as Gaussian8, from range [1.8, 3.2], [1.35, 2.40] and [0.80, 1.60] for scale factors 4, 3 and 2 respectively‘’