✏️Title:
#Unsupervised #Representation Learning with #Deep #Convolutional #Generative #Adversarial Networks
✏️abstract:
In recent years, supervised learning with convolutional networks (CNNs) has seen huge adoption in computer vision applications. Comparatively, unsupervised learning with CNNs has received less attention. In this work we hope to help bridge the gap between the success of CNNs for supervised learning and unsupervised learning. We introduce a class of CNNs called deep convolutional generative adversarial networks (DCGANs), that have certain architectural constraints, and demonstrate that they are a strong candidate for unsupervised learning. Training on various image datasets, we show convincing evidence that our deep convolutional adversarial pair learns a hierarchy of representations from object parts to scenes in both the generator and discriminator. Additionally, we use the learned features for novel tasks - demonstrating their applicability as general image representations.

🔗https://arxiv.org/pdf/1511.06434v2.pdf

"Under review as a conference paper at ICLR 2016"

مقاله‌ی جالب کمپانی اپل!
( Submitted for review to a conference on Nov 15, 2016)

✏️Title:
Learning from Simulated and Unsupervised Images through Adversarial Training

✏️abstract:
With recent progress in graphics, it has become more tractable to train models on #synthetic images, potentially avoiding the need for expensive annotations. However, learning from synthetic images may not achieve the desired performance due to a gap between synthetic and real image distributions. To reduce this gap, we propose Simulated+Unsupervised (S+U) learning, where the task is to learn a model to improve the realism of a simulator's output using #unlabeled real data, while preserving the annotation information from the simulator. We develop a method for S+U learning that uses an #adversarial network similar to #Generative Adversarial Networks (#GANs), but with synthetic images as inputs instead of random vectors. We make several key modifications to the standard GAN algorithm to preserve annotations, avoid artifacts and stabilize training: (i) a 'self-regularization' term, (ii) a local adversarial loss, and (iii) updating the discriminator using a history of refined images. We show that this enables generation of highly realistic images, which we demonstrate both qualitatively and with a user study. We quantitatively evaluate the generated images by training models for gaze estimation and hand pose estimation. We show a significant improvement over using synthetic images, and achieve state-of-the-art results on the MPIIGaze dataset without any labeled real data.

🔗https://arxiv.org/abs/1612.07828v1
🔗https://arxiv.org/pdf/1612.07828v1.pdf

#unlabeled_data #unsupervised #unsupervised_learning #Generative #Generative_Models

Generative Adversarial Denoising Autoencoder for #Face Completion
pic: http://www.cc.gatech.edu/~hays/7476/projects/Avery_Wenchen/images/one.png

🔗http://www.cc.gatech.edu/~hays/7476/projects/Avery_Wenchen/
#GAN
#Generative #adversarial #Generative_Models #Autoencoder

Image-to-Image Translation in #Tensorflow


در این کار شبکه های شرطی در مقابل حریف آموزش دیده اند که یک نگاشت از تصویر ورودی به تصویر خروجی بیابند به نحوی که بتواند با لبه ها به عنوان وردی، اشیاء و تصویر شبیه به واقعی بازسازی کند.

✒️ Image-to-Image Translation with Conditional Adversarial Nets
🔗 https://phillipi.github.io/pix2pix/

✒️Image-to-Image Demo:
🔗 https://affinelayer.com/pixsrv/

✒️Tensorflow port of Image-to-Image Translation with Conditional Adversarial Nets
🔗https://github.com/affinelayer/pix2pix-tensorflow


#conditional #adversarial networks

Generating Videos with Scene Dynamics
video: http://bit.ly/2q6THM9

تبدیل تصویر به فیلم.
هوش مصنوعی ای که قادر است تنها با یک تصویر ثابت، فیلم چند ثانیه ای حاوی حرکت خروجی دهد...

در این روش به صورت بدون ناظر دو سال ویدیوی جمع آوری از فلیکر به شبکه آموزش داده شده است، سپس شبکه توانسته که نگاشتی از تصاویر به فیلم های چند ثانیه ای ایجاد کند.

🔗 http://web.mit.edu/vondrick/tinyvideo/

#generative #adversarial #GAN #deep_learning

#مقاله
✔️ایجاد یک نگاشت از تصور به تصویر:
در این کار شبکه های شرطی در مقابل حریف (GAN) آموزش دیده اند که یک نگاشت از تصویر ورودی به تصویر خروجی بیابند...


Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
[UC Berkeley] pic: http://bit.ly/2s2OTsm

🔗abstract:
https://arxiv.org/abs/1703.10593

🔗Paper:
https://arxiv.org/pdf/1703.10593.pdf

🔗Project Page:
https://junyanz.github.io/CycleGAN/

🔗codes:
https://github.com/junyanz/CycleGAN
https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix


Our goal is to learn a mapping G: X → Y such that the distribution of images from G(X) is indistinguishable from the distribution Y using an adversarial loss.

مرتیط به مقاله ی:
https://t.me/cvision/171

#CycleGAN #GAN #Generative #CNN #Convolutional #deep_learning #adversarial #Generative_Models #Generative

تبدیل اسب به گورخر!
ایجاد نگاشت تصویر به تصویر توسط هوش مصنوعی...

اطلاعات بیشتر:
https://t.me/cvision/214

#CycleGAN #GAN #Generative #CNN #Convolutional #deep_learning #adversarial #generative

#آموزش در سایت ماکروسافت
Learning Image to Image Translation with CycleGANs
[Published June 12, 2017]
https://www.microsoft.com/reallifecode/2017/06/12/learning-image-image-translation-cyclegans/

مرتبط با https://t.me/cvision/214

#CycleGAN #GAN #Generative #CNN #Convolutional #deep_learning #adversarial #Generative_Models #Generative

#مقاله
Perceptual Generative #Adversarial Networks for Small Object Detection

https://arxiv.org/abs/1706.05274

#مقاله #سورس_کد
StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation
+ PyTorch Implementation of StarGAN - #CVPR_2018


🔗abstract:
https://arxiv.org/abs/1711.09020
🔗Paper:
https://arxiv.org/pdf/1711.09020.pdf
🔗Code:
https://github.com/yunjey/StarGAN
🎬Video Demo:
https://www.youtube.com/watch?v=EYjdLppmERE


#GAN #stargan #pytorch #generative #adversarial

StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation
+ کد
اطلاعات بیشتر

https://t.me/cvision/646

#GAN #stargan #pytorch #generative #adversarial

#سورس_کد

[pic: https://t.me/cvision/870]

Image-to-Image Translation in #Tensorflow_js

Fast image-to-image translation in the #browser.
With 3 new trained models .
+ processed dataset of 1000 images for edges2cats translation.

Demo: https://zaidalyafeai.github.io/pix2pix/cats.html
Code(Tensorflow.js): https://github.com/zaidalyafeai/zaidalyafeai.github.io/tree/master/pix2pix

مرتبط با:
https://t.me/cvision/863
https://t.me/cvision/171
https://t.me/cvision/170

#conditional #adversarial #GAN
#pix2pix

#مقاله

One pixel attack for fooling deep neural networks

در این مقاله تنها با تغییر یک پیکسل از تصویر موفق شده adversarial attack انجام بده وشبکه را فریب بده!

https://arxiv.org/abs/1710.08864

#adversarial

#مقاله
فریب شبکه های object detection در تشخیص افراد.
Fooling automated surveillance cameras: adversarial patches to attack person detection. https://arxiv.org/abs/1904.08653

🙏Thanks to: @vahidreza01
#adversarial #gan

#سورس_کد #مقاله

در این کار پچ هایی ایجاد و چاپ کردند که با وجود آنها الگوریتم شناسایی افراد روی شما کار نخواهد کرد و الگوریتم فریب خواهد خورد.
Fooling automated surveillance cameras: #adversarial patches to attack person detection

مقاله:
https://arxiv.org/abs/1904.08653

سورس کد:
https://gitlab.com/EAVISE/adversarial-yolo

#adversarial #adversarial_attack #person_detection