GAN papers

本文提出了一种新的生成对抗网络(GANs)生成器架构,借鉴了风格迁移领域的思想。该架构能自动学习高阶属性(如面部姿态和身份)与图像中的随机变化(如雀斑和头发)之间的无监督分离,并允许对合成过程进行直观且特定尺度的控制。新生成器在传统分布质量指标上提升了state-of-the-art,并展现出更好的插值性质和潜在变量的解耦效果。为量化插值质量和解耦,我们提出了两个适用于任何生成器架构的新自动化方法。此外,我们还发布了一个包含高度多样化和高质量人脸的大型数据集。
摘要由CSDN通过智能技术生成

A Style-Based Generator Architecture for Generative Adversarial Networks

CVPR 2019 NVIDIA https://arxiv.org/abs/1812.04948
We propose an alternative generator architecture for generative adversarial networks, borrowing from style transfer literature. The new architecture leads to an automatically learned, unsupervised separation of high-level attributes (e.g., pose and identity when trained on human faces) and stochastic variation in the generated images (e.g., freckles, hair), and it enables intuitive, scale-specific control of the synthesis. The new generator improves the state-of-the-art in terms of traditional distribution quality metrics, leads to demonstrably better interpolation properties, and also better disentangles the latent factors of variation. To quantify interpolation quality and disentanglement, we propose two new, automated methods that are applicable to any generator architecture. Finally, we introduce a new, highly varied and high-quality dataset of human faces.
在这里插入图片描述

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值