Convolutional Autoencoder based Denoising of images corrupted with synthetic noise. This is based on a Kaggle Challenge. The data set is also available in the repository itself. Train_cleaned and Train constitute the ground truth and corrupted dataset respectively. Test folder contains the test dataset.
SamarjeetKaur/Denoising-Dirty-Documents-Kaggle
Convolutional Autoencoder based Denoising of images corrupted with synthetic noise. This is based on a Kaggle Challenge.
Jupyter Notebook