The paper review of CycleGANs where we use CycleGANs for generating Ukiyo-e art. Image-to-image translation involves generating a new synthetic version of a given image with a specific modification. The CycleGAN is a technique that involves the automatic training of image-to-image translation models. The models are trained in an unsupervised manner using a collection of images from the source and target domain that do not need to be related in any way. The core idea was to build and understand a CycleGAN to generate Ukiyo-e art.
The repository has 3 files:
- code.py
- prep_data.py
- CycleGANs.ipynb This file is the whole composed code including the data preparation and model definitions along with code explanation and paper review.
The dataset is also made available in the ukiyoe2photo folder. The generated images/plots are stored in Results folder.