Use PCA to calculate a reduced number of "principal components" (aka eigenvectors) of 64x64 anime faces, the playground let's you mess with the values of these principal components and generate new anime faces.
Dataset used : https://www.kaggle.com/soumikrakshit/anime-faces
Read up on the whole idea : https://en.wikipedia.org/wiki/Eigenface
Libraries :
- sklearn
- joblib
- flask
- flask_bootstrap