85a - What are Autoencoders and what are they used for?

Autoencoders are artificial neural networks that can be trained to encode data (e.g. images) into a compressed representation and then trained to reconstruct the encoded data back to its original shape (& size).
Autoencoders can be designed for various applications including, denoising, image colorization, and anomaly detection. This tutorial provides and explanation of autoencoders and their uses.
The code from this video is available at: github.com/bnsreenu/python_fo...

[@saigopal4361]

Just an addition, at 5:44, the reason why the output image will not "perfectly" be same as the input image is that the code space has fewer dimensions compared to the input space. The reason for choosing the code space to have fewer dimensions than the input space is to avoid duplicating the input which is a simple way to achieve the task that we are discussing.

[@erfanelmtalab1615]

sir you are literally a life saver , thank you so much ...

[@adithiajo5028]

very nice material thanks,im waiting denoising with autoencoder :)

[@caiyu538]

GREAT LECTURES

[@TruhMann]

Thank you!

[@amirhosseintalebi6770]

thank you Sir

[@user-vy5sg2lp1w]

its amazing video

[@user-vo7iz5bx6j]

Thank you.

[@DigitalSreeni]

You're welcome!

[@Mach89]

Isn’t instance segmentation with unet example of domain adaptation? It maps image into segments using CNN. 🤔

[@EB3103]

Thank you!

[@DigitalSreeni]

You're welcome!