BriefGPT.xyz
Jul, 2018
我的 GAN 有多强？
How good is my GAN?
HTML
PDF
Konstantin Shmelkov, Cordelia Schmid, Karteek Alahari
TL;DR
本研究提出了两种基于图像分类的测量标准，GAN-train和GAN-test，以评估各种GAN方法的召回率和准确率。研究结果表明，数据集的难度与GAN质量呈负相关，并发现这两种度量标准可以更好地评估GAN的性能。
Abstract
generative adversarial networks
(
gans
) are one of the most popular methods for generating images today. While impressive results have been validated by visual inspection, a number of quantitative criteria have em
→