Keras implementation of neural style transfer [Gatys et. al. 2015]
Neural style transfer is a method of transferring characteristics of an image to another image using a pretrained deep convolutional neural network like VGG16. Essentially same priciples used by popular mobile app Prisma.
An example of Taj Mahal painted in the style of Van Gogh's "Starry Night".
Another example of transferring the characteristics of Picasso's painiting to an image of a ship.
Make sure you have installed keras and pillow before running the python file. Use pip install Pillow.
Refernces: 1)Andrew Ng's coursera lectures - https://www.youtube.com/watch?v=R39tWYYKNcI&list=PLkDaE6sCZn6Gl29AoE31iwdVwSG-KnDzF&index=37
2)Detailed explanation of the code - https://harishnarayanan.org/writing/artistic-style-transfer/