Click the card below to follow the " CVer " public account
AI/CV heavy dry goods, delivered as soon as possible
Reprinted from: face and human body reconstruction
Comics, cartoons, caricatures, Japanese animation, Pixar animation, no matter what style you have, you can easily reproduce it!
The author of this article is from Nanyang Technological University. The author proposes DualStyleGAN to do high-resolution (1024x1024) portrait style transfer based on reference samples. DualStyleGAN can flexibly control the two styles of the original face domain and the artistic portrait domain, which is different from StyleGAN. Characterize the content and style of portraits by using the intrinsic style path and the newly introduced extrinsic style path to provide more natural style transfer. The author's well-designed extrinsic style path enables DualStyleGAN to hierarchically model colors and complex structural styles, thereby accurately replicating the style of reference examples. In addition, DualStyleGAN is very efficient on training data, with only about 200 images , it can be trained with good results,
Homepage: https://www.mmlab-ntu.com/project/dualstylegan
Code: https://github.com/williamyang1991/DualStyleGAN
Click "Read the original text" below to go directly to the GitHub code!
▲Cartoon style migration▲
▲Caricature style migration▲
▲Japanese animation style migration▲
The image below shows more face image style transfer results!
▲Cartoon style▲
▲Caricature style▲
▲Japanese animation style▲
▲Pixar animation style▲
▲Manga style▲
▲Slam Dunk Style▲
The images below show the results of DualStyleGAN compared to current State-of-the-art methods. The objects compared by the author include 6 methods, including image-to-image-translation-based StarGAN2 , GNR , U-GAT-IT , and StyleGAN-based UI2I-style , Toonify , Few-Shot Adaptation ( FS-Ada ), the results Showing that DualStyleGAN can transfer the color and result information from the reference image well!
1. Pastiche Master: Exemplar-Based High-Resolution Portrait Style Transfer. Shuai Yang, Liming Jiang, Ziwei Liu, Chen Change Loy. CVPR, 2022.
ICCV和CVPR 2021论文和代码下载
后台回复：CVPR2021，即可下载CVPR 2021论文和代码开源的论文合集
后台回复：ICCV2021，即可下载ICCV 2021论文和代码开源的论文合集
后台回复：Transformer综述，即可下载最新的3篇Transformer综述PDF
CVer-GAN交流群成立
扫描下方二维码，或者添加微信：CVer6666，即可添加CVer小助手微信，便可申请加入CVer-Transformer 微信交流群。另外其他垂直方向已涵盖：目标检测、图像分割、目标跟踪、人脸检测&识别、OCR、姿态估计、超分辨率、SLAM、医疗影像、Re-ID、GAN、NAS、深度估计、自动驾驶、强化学习、车道线检测、模型剪枝&压缩、去噪、去雾、去雨、风格迁移、遥感图像、行为识别、视频理解、图像融合、图像检索、论文投稿&交流、PyTorch、TensorFlow和Transformer等。
一定要备注：研究方向+地点+学校/公司+昵称（如GAN+上海+上交+卡卡），根据格式备注，可更快被通过且邀请进群
▲扫码或加微信: CVer6666，进交流群
CVer学术交流群（知识星球）来了！想要了解最新最快最好的CV/DL/ML论文速递、优质开源项目、学习教程和实战训练等资料，欢迎扫描下方二维码，加入CVer学术交流群，已汇集数千人！
▲扫码进群
▲点击上方卡片，关注CVer公众号
整理不易，请点赞和在看