Lesson 14: Deep Learning Part 2 2018 - Super resolution; Image segmentation with Unet

preview_player
Показать описание

In this final lesson, we do a deep dive into super resolution, an amazing technique that allows us to restore high resolution detail in our images, based on a convolutional neural network. In the process, we’ll look at a few modern techniques for faster and more reliable training of generative convnets.

We close with a look at image segmentation, in particular using the Unet architecture, a state of the art technique that has won many Kaggle competitions and is widely used in industry. Image segmentation models allow us to precisely classify every part of an image, right down to pixel level.

We hope you enjoyed your deep learning journey with us! Now that you’ve finished, be sure to drop by the forums to tell us how you’re using deep learning in your life or work, or what projects you’re considering working on now.
Рекомендации по теме
Комментарии
Автор

Segmentation: 1:29:11
UNet: 1:44:33

Huge thanks Jeremy, this is the most thorough UNet implementation on YouTube

patrickjdarrow
Автор

The best online teacher I've ever had.

adhoc
Автор

Thanks. You are a great teacher Jeremy.

TASumesh
Автор

Can someone tell me the difference between annotated images, labels, masks and truth images?

tanmaydeshpande
Автор

Jeremy's awesome! thank you for this

elzilcho
Автор

Is it possible to perform superresolution without changing the resolution, but just add details?

santhapurharsha
Автор

Can I ask the size that is allocated in gpu for model training for 1024x1024 batch size = 4?

In my case 400X400 batchsize 14 takes almost 11gb.

faruknane
Автор

I need to segment weed plant images can I get UNet in matlab to apply it on my dataset?

ahmedhusham
Автор

Is it necessary for me to go through all 14 lectures to understand this super resolution lecture?
Or is it fine if i directly go through lesson 14. ( I do not have much knowledge on ip, neural network, ai or deep learning)

globalhack
Автор

very scattered lecture.Need to more compact like stanford or Prof nando

amitabhachakraborty
Автор

Please Give me the link to this Notebook Or Source code

mockingbird