247 - Conditional GANs and their applications

Conditional Generative Adversarial Network cGAN
A GAN model generates a random image from the domain.
The relationship between points in the latent space and the generated images is hard to map.
A GAN can be trained so that both the generator and the discriminator models are conditioned on the class label (or other modalities).
As a result, the trained generator model can be used to generate images of a given type using the class label (or another condition).
GAN can be conditioned using other image modalities (image to image translation).
The conditioning is performed by feeding the class label into both the discriminator and generator as an additional input layer.
A few applications:
Image-to-Image Translation: Pix2Pix GAN
CycleGAN: Transform images from one set into images that could belong to another set.
Super-resolution: Increase the resolution of images, adding detail where necessary to fill in blurry areas.
Text-to-Image Synthesis: Take text as input and produce images as described by the text.

[@cplusplus-python]

So excited to get to the Code part of GAN, thanks Prof.

[@hanemellethy9036]

Great tutorial. Very simple and informative video. I really appreciate your easy and helpful way of explanation. Thanks a million.

[@rubenguerrerorivera7462]

So beautifully explained, so smooth and highly enjoyable! Thanks a lot Dr.

[@rahchandr]

Sreeni sir, great going, these sessions are profoundly useful.

[@deepalisharma1327]

Amazing work, really appreciate your efforts. 🙏🏻 Please keep making such videos.

[@hessame9496]

I cannot thank you enough for sharing your knowledge and preparing and publishing these great tutorials.

[@DigitalSreeni]

Thank you :)

[@prashantjaiswal5260]

Your videos are genuinely knowledgable sir ...Keep providing with such great contents .
Please provide these slides also if possible

[@Tiago_R_Ribeiro]

You are very good and very patient teacher. I watch your videos every single day. Thanks for making videos for mere mortals like me! :D

[@wessauder7708]

Thank you so much for this detailed and easy to follow demonstration! It's a major component of my grad research and you have tied the concepts together so well that it really complements and reinforces my understanding.

[@calvinchristiano7119]

Thankyou sir for this amazing tutorial, very clear explanation, very patient teacher....i really appreciate that. Stay healthy sir

[@ciancian2861]

Thank you!! I'm a data science student and I will start my thesis on this topic next week. Great introduction.

[@DigitalSreeni]

Best of luck!

[@vafakahnamoie29]

Sir, your tutorials make confusing and complicated AI topics to easy and comprehensible concepts for us. Thanks a lot professor

[@DigitalSreeni]

You are most welcome

[@eli_m6556]

Great vid as always. Your videos are great to watch even if I’m not working on the given topic.

[@DigitalSreeni]

If you only watch videos on topics that you relate to, then how do you learn about other topics? I think it is very important for us to learn about various topics so we can find the one that really interests us.

[@parthmodi5734]

You deserve a huge round of applause, Thanks for this great content. God bless you:)

[@karamsahoo6852]

Yes sir please make more videos on different GAN architectures.

[@TheCzarsoham]

Thank you for the clear explanation! I really appreciate your videos

[@DigitalSreeni]

Please keep watching :)

[@mohamednaji6178]

thank you so much you are amazing I have learned so much from you

[@cheemay]

Great Informative video. Now understand conditional GAN. Thanks #DigitalScreeni
Waiting For StackGan Implementation

[@mohitmeena541]

worth every second. thanks a lot!

[@aashwinsharma1859]

Thnx a lot for the wonderful explanation

[@emmanuellabudu9524]

Awesome video. Thank you.

[@yanivnoema6848]

Thank you for this video!

[@umairsabir6686]

Hi Sreeni,
You were great as always. Do you have Mask RCNN using TF2 in your roadmap or not ?

[@mtahirrasheed5538]

Can you make videos on the transformers? Vision Transformer for the classification. The main issue is in understanding the input/output shape, number of patches for different images sizes etc. Thanks in advance.

[@Suman-zm7wx]

Finally you are back in the game sir 💚💚

[@rajeevgoel6170]

Very Good Explained Sir

[@aqibmumtaz1262]

Excellent Great video sir

[@khalidelamraoui256]

thank you for the effort , can i ask you to make an applications for ESRGAN to understand it very well

[@mariovrpereira]

very good, thank you

[@javlontursunov6527]

Thank you so much :)

[@alanpress3019]

Sir, Thank you so much. Are you planning to do some tutorials on meat-learning in the future, e.g., learning to learn gradient descent by gradient descent, or learning to learn without gradient descent by gradient descent, and keras implementation?

[@PUBUDUCG]

Good information . . .

[@ameerazam3269]

Thanks you Sir ... UOH love ..

[@borundas6443]

Sir, do you have any video how to make images from text using GANs? I really need some good tutorial on that.

[@dannyboi7182]

thank you sir

[@profesorjan7614]

Excellent explanation!!!!!!!! Thanks!

[@DigitalSreeni]

Glad it was helpful!

[@botirkarim9293]

Keep continue good luck!

[@fatmagulkurt2080]

How can I apply k-fold cross validation in the 195. tutorial(195 - Image classification using XGBoost and VGG16 imagenet as feature extractor). I wish you may help me in this situation. Because the most common problem in practice is overfittig. How can I overcome this in this code Thank you for all your effort Sir.

[@mustajabhussain9167]

Very helpful video. Can you please tell me that can we perform semantic segmentation using conditional GAN. In this video, you talk about getting real image from semantic segmented image. But can we perform the task we did using UNet architecture (getting semantic segmented mask of specific image)

[@RohanPaul-AI]

Hi Mustajab - Stumbled upon your comment, and I think this paper did what you are talking about - arxiv.org/abs/1708.05227
They used conditional GAN and train a semantic segmentation CNN along with an adversarial network that discriminates segmentation maps coming from the ground truth or from the segmentation network for BraTS 2017 segmentation task
More specifically, they used patient-wise ”U-Net” as a generator and ”Markovian GAN” as an discriminator.

[@Darshan0070]

How to match images for similar products??

[@user-ec9et8nn9l]

Is there a video that can help me with binarization using GAN so i can watch that one

[@gunjannaik7575]

How to randomize the number of images that are passed in each epoch?

[@avinolia]

can we use GANs or CGANs to balance the dataset? Please explain sir

[@prashantpaikrao9534]

Sir, how we can use GAN for noise removal in document images?

[@anonnozahir7375]

Sir can I use this code for doing RGB to Grayscale images?

[@rs9130]

Anything is possible and everything is easy with DIgital Sreeni

[@mekuriawayelegn1598]

Thanks sir.

[@DigitalSreeni]

Most welcome

[@prashantpaikrao9534]

Sir, do you have made any video on deep dense GAN? If yes please send me it's lesson number or link... 🙏🏼🙏🏼🙏🏼

[@sajins9708]

Sir, would you please upload tutorials on object detection algorithms like faster RCNN and fast RCNN.

[@DigitalSreeni]

Sometime in future but definitely not in the next couple of months. Thanks for the suggestion though, I need to find time to put together code that works and then plan videos. Takes time.

[@caiyu538]

Thumb up your video though it is busy for something else recently.

[@saritagautam3483]

subscribed 🤙

[@DigitalSreeni]

Thanks :)

[@furry2fun]

sir can you please share these slides

[@ssangeetha7497]

Really I got interest in deep learning methods on watching ur tutorials.sir I wish to clarify doubts in my deep learning based work . So can you share your email I'd.