This is an unofficial implementation for the paper "NEIGHBOR2NEIGHBOR: SELF-SUPERVISED DENOISING FROM SINGLE NOISY IMAGES"
This is an implementation for the paper "NEIGHBOR2NEIGHBOR"https://arxiv.org/abs/2101.02824, and i have tied this method for real noise removal task in my own dataset, which has presented some effects to some degree. But i haven't applied it to some traditional denoising tasks, e.g., Gaussian nosie and Posson noise removing. You could try it easily in this codes.
- Pytorch > 1.3.0
- Nvidia apex ( this codes are easily used for multi-gpus training)
- Some Python packages
- First, you need to add the data preprocessing file based your tasks.
- Update the config.yaml file
- training and testing with
python main.py
This is an extended work for Noise2Noise https://arxiv.org/abs/1803.04189, which all depend on the powerful zero-mean noise prior hypothesis.