Have you wondered how your nearest café or the streets would look like if it were in an anime? How would you like as an anime character? Relieve that experience using AnimeGAN.
This is my Computer Vision final project that discusses the widely acclaimed General Adversarial Network (GAN) : AnimeGAN that uses machine learning through neural style transfer and generative adversarial networks (GANs) to achieve a real "anime-like" conversion of an image.
I utilized the COCO dataset to generate 3 anime styles supported via AnimeGAN v3 to achieve some amazing results. You can try the sample on a smaller dataset.
https://github.com/TachibanaYoshino/AnimeGANv3 https://github.com/TachibanaYoshino/AnimeGAN/blob/master/doc/Chen2020_Chapter_AnimeGAN.pdf https://s3.amazonaws.com/fast-ai-coco/val2017.zip