DD2402 Advanced Individual Course in Computational Biology Project.
Implementation of Osokin et al. GANs for Biological Image Synthesis (link). In this project we try to (at least partially) implement the work presented in the paper. We also extend the original work by testing smaller models and also using statistical testing to support our claim that one-class GANs perform superiorly to multi-class ones when it comes to the fidelity of the generated cellular evolution images.
For a more in-depth analysis of the related concepts and the followed methodology, please consult the final report pdf. This can also be found in the repo root, biogans-report-thanos-charisoudis.pdf.