Skip to content

Latest commit

 

History

History
executable file
·
34 lines (20 loc) · 1007 Bytes

README.md

File metadata and controls

executable file
·
34 lines (20 loc) · 1007 Bytes

Image Restoration: denoising and splitting

Welcome to the Image Restoration exercises. In this part of the course, we will explore how to use deep learning to denoise images, with examples of widely used algorithm for both supervised and unsupervised denoising. We will also explore the difference between unstructured and structured noise, or between UNet (which you are familiar with by now) and VAE architectures (see COSDD exercise)!

Finally, we have bonus exercises for those wanted to explore more denoising algorithms or image splitting!

Setup

Please run the setup script to create the environment for these exercises and download data.

source setup.sh

Exercises

  1. Context-aware restoration
  2. Noise2Void
  3. Correlated and Signal Dependent Denoising (COSDD)
  4. DenoiSplit

Bonus