The goal of the mini-projects is to implement a Noise2Noise model. A Noise2Noise model is an image denoising network trained without a clean reference image. The original paper can be found at https://arxiv.org/abs/1803.04189.
Mini-project 1 is implemented using the standard PyTorch framework and in Mini-project 2, we had to write our own blocks.
See the EE559_2022ProjDescription.pdf for more details.