社区所有版块导航
Python
python开源   Django   Python   DjangoApp   pycharm  
DATA
docker   Elasticsearch  
aigc
aigc   chatgpt  
WEB开发
linux   MongoDB   Redis   DATABASE   NGINX   其他Web框架   web工具   zookeeper   tornado   NoSql   Bootstrap   js   peewee   Git   bottle   IE   MQ   Jquery  
机器学习
机器学习算法  
Python88.com
反馈   公告   社区推广  
产品
短视频  
印度
印度  
Py学习  »  Git

Image-to-Image的论文汇总(含github代码)

极市平台 • 6 年前 • 848 次点击  

极市平台(ExtremeMart)是深圳极视角旗下的专业视觉算法开发与分发平台,为开发者提供行业场景集,每月上百真实项目需求,算法分发,技术共享等,旨在联合开发者建立起良好的计算机视觉生态。已与上百名开发者建立了合作并转化了上百种视觉算法。

PS.本周四(11月15日)晚,TEE首席架构师、TEE AI Lab资深研究员邓文彬将为我们讲解如何在GPU/CPU/移动端高效训练CNN网络,公众号回复“35”即可获取直播详情。


图像生成一直是计算机视觉领域非常有意思的方向,图像到图像的变换是其中一个非常重要的应用,使用图像到图像的变换,可以完成非常多有趣的应用,可以把黑熊变成熊猫,把你的照片换成别人的表情,还可以把普通的照片变成毕加索风格的油画,自从GAN横空出世之后,这方面的应用也越来越多,下面是对这个领域的相关论文的一个整理,而且大部分都有代码!

github地址:https://github.com/ExtremeMart/image-to-image-papers

这是一个图像到图像的论文的汇总。

论文按照arXiv上第一次提交时间排序。


监督学习

NoteModelPaperConferencepaper link(arXivcode link(github)

pix2pixImage-to-Image Translation with Conditional Adversarial NetworksCVPR 20171611.07004junyanz/pytorch-CycleGAN-and-pix2pix

Contextual GANImage Generation from Sketch Constraint Using Contextual GANECCV 20181711.08972

pix2pix-HDHigh-Resolution Image Synthesis and Semantic Manipulation with Conditional GANsCVPR 20181711.11585NVIDIA/pix2pixHD
一对多BicycleGANToward Multimodal Image-to-Image TranslationNIPS 20171711.11586junyanz/BicycleGAN

contour2imSmart, Sparse Contours to Represent and Edit Images CVPR 20181712.08232website
分离Cross-domain disentanglement networksImage-to-image translation for cross-domain disentanglementNIPS 20181805.09730


非监督学习


非监督学习- 通用

NoteModelPaperConferencepaper link(arXivcode link(github)

DTNUnsupervised Cross-Domain Image GenerationICLR 20171611.02200 yunjey/domain-transfer-network (unofficial)

UNITUnsupervised image-to-image translation networksNIPS 20171703.00848mingyuliutw/UNIT

DiscoGANLearning to Discover Cross-Domain Relations with Generative Adversarial NetworksICML 20171703.05192SKTBrain/DiscoGAN

CycleGANUnpaired Image-to-Image Translation using Cycle-Consistent Adversarial NetworksICCV 20171703.10593junyanz/pytorch-CycleGAN-and-pix2pix

DualGANDualGAN: Unsupervised Dual Learning for Image-to-Image TranslationICCV 20171704.02510duxingren14/DualGAN

DistanceGANOne-Sided Unsupervised Domain MappingNIPS 20171706.00826sagiebenaim/DistanceGAN

Triangle GANTriangle Generative Adversarial NetworksNIPS 20171709.06548LiqunChen0606/Triangle-GAN
特征点导向G2-GANGeometry Guided Adversarial Facial Expression SynthesisMM 20181712.03474

CartoonGANCartoonGAN: Generative Adversarial Networks for Photo CartoonizationCVPR 2018thecvfunofficial test, unofficial pytorch
非对抗NAM NAM: Non-Adversarial Unsupervised Domain MappingECCV 20181806.00804facebookresearch/nam

SCANUnsupervised Image-to-Image Translation with Stacked Cycle-Consistent Adversarial NetworksECCV 20181807.08536
空洞卷积,提高形状的变形GANimorphImproved Shape Deformation in Unsupervised Image to Image TranslationECCV 20181808.04325brownvc/ganimorph
实例感知InstaGANInstance-aware image-to-image translationICLR 2019 (in review)openreview


非监督学习- 注意力机制或者模板导向机制

NoteModelPaperConferencepaper link(arXivcode link(github)
模板ContrastGANGenerative Semantic Manipulation with Mask-Contrasting GANECCV 20181708.00315
注意力机制DA-GANDA-GAN: Instance-level Image Translation by Deep Attention Generative Adversarial NetworksCVPR 20181802.06454
模板/ 注意力Attention-GANAttention-GAN for Object Transfiguration in Wild Images
1803.06798
注意力Attention guided GANUnsupervised Attention-guided Image to Image TranslationNIPS 20181806.02311AlamiMejjati/Unsupervised-Attention-guided-Image-to-Image-Translation
注意力, 单边
Show, Attend and Translate: Unsupervised Image Translation with Self-Regularization and Attention
1806.06195


非监督学习-多对多(属性)

NoteModelPaperConferencepaper link(arXivcode link(github)

Conditional CycleGANConditional CycleGAN for Attribute Guided Face Image Generation ECCV 20181705.09966

StarGANStarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image TranslationCVPR 20181711.09020yunjey/StarGAN

AttGANAttGAN: Facial Attribute Editing by Only Changing What You Want
1711.10678LynnHo/AttGAN-Tensorflow

ComboGANComboGAN: Unrestrained Scalability for Image Domain TranslationCVPRW 20181712.06909AAnoosheh/ComboGAN

AugCGAN (Augmented CycleGAN)Augmented CycleGAN: Learning Many-to-Many Mappings from Unpaired DataICML 20181802.10151aalmah/augmented_cyclegan

SG-GANSparsely Grouped Multi-task Generative Adversarial Networks for Facial Attribute ManipulationMM 20181805.07509zhangqianhui/Sparsely-Grouped-GAN

GANimationGANimation: Anatomically-aware Facial Animation from a Single ImageECCV 2018 (honorable mention)1807.09251albertpumarola/GANimation


非监督学习- 分离(与/或样本导向)

NoteModelPaperConferencepaper link(arXivcode link(github)
非分离, 纹理导向TextureGAN TextureGAN: Controlling Deep Image Synthesis with Texture PatchesCVPR 20181706.02823janesjanes/Pytorch-TextureGAN

XGANXGAN: Unsupervised Image-to-Image Translation for Many-to-Many MappingsICML 20181711.05139dataset

ELEGANTELEGANT: Exchanging Latent Encodings with GAN for Transferring Multiple Face AttributesECCV 20181803.10562Prinsphield/ELEGANT

MUNITMultimodal Unsupervised Image-to-Image TranslationECCV 20181804.04732NVlabs/MUNIT

cd-GAN (Conditional DualGAN)Conditional Image-to-Image TranslationCVPR 2018 1805.00251

EG-UNITExemplar Guided Unsupervised Image-to-Image Translation
1805.11145

DRITDiverse Image-to-Image Translation via Disentangled RepresentationsECCV 20181808.00948HsinYingLee/DRIT
分分离, 人脸化妆导向BeautyGANBeautyGAN: Instance-level Facial Makeup Transfer with Deep Generative Adversarial NetworkMM 2018author

UFDNA Unified Feature Disentangler for Multi-Domain Image Translation and ManipulationNIPS 20181809.01361Alexander-H-Liu/UFDN


本文选自github

作者:lzhbrian

编译:ronghuaiyang
来源:

https://mp.weixin.qq.com/s/KiIpZb-9vq9bagcLV1aXJQ




*推荐阅读*

DeepMind&VGG提出基于集合的人脸识别算法GhostVLAD,精度远超IJB-B 数据集state-of-the-art

资源 | 谷歌开源AdaNet:基于TensorFlow的AutoML框架

微软开源的深度学习模型转换工具MMdnn


PS.本周四(11月15日)晚,TEE首席架构师、TEE AI Lab资深研究员邓文彬将为我们讲解如何在GPU/CPU/移动端高效训练CNN网络,公众号回复“35”即可获取直播详情。左下角阅读原文查看更多直播预告。



今天看啥 - 高品质阅读平台
本文地址:http://www.jintiankansha.me/t/lPf64ngKso
Python社区是高质量的Python/Django开发社区
本文地址:http://www.python88.com/topic/26368
 
848 次点击