Black Swans


2 points



Convolution Transpose and Sub-pixel Convolution

I read an interesting paper called Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network.
I am having some difficulty understanding the concept of "Pixel Shuffle", which uses sub-pixel convolution to upscale the input images in the final layer.

If anybody could explain how exactly the sub-pixel and transpose convolution operations are performed, I'll be very thankful.

P.S. I referred to this document too, but it didn't help much.

Hey there!

I am Utkarsh Verma, a senior undergrad at Delhi Technological University, Delhi, India. I just got to know about Machine Learning/AI recently and I am taking the Udacity course on deep learning using pytorch currently. I am liking it this far, and hope to collaborate with some of the members to work on some interesting projects and to get to know about the opportunities around the globe.