In this project we use should use the MNIST datset and add noise to it then we should train a model to denoise them.
The following things have been done in this project
- Download the MNIST dataset and create a dataloader that adds gaussian noise to the input images.
- Design and train an AutoEncoder on the MNIST dataset to denoise the noisy images.
- Visualize original images, their corresponding noisy images and their reconstructed versions side by side.
- Repeat the previous steps using PCA algorithm.
This project was the first part of third phase of Intoduction to machine learning course which is lectured at electrical engineering(EE) department of Sharif university of technology(SUT)