BriefGPT.xyz
Mar, 2023
CAP-VSTNet: 内容关联性保持的通用风格迁移
CAP-VSTNet: Content Affinity Preserved Versatile Style Transfer
HTML
PDF
Linfeng Wen, Chengying Gao, Changqing Zou
TL;DR
本文提出了一种名为CAP-VSTNet的新框架，其中包括一个新的可逆残差网络和一个无偏线性变换模块，用于各种样式转换。通过Matting Laplacian训练损失，该框架可解决由线性变换引起的像素亲和力丢失问题，避免引入冗余信息并提高风格化效果。
Abstract
content affinity loss
including feature and pixel affinity is a main problem which leads to artifacts in
photorealistic
and
video style transfer<
→