Recently, I have seen many beautiful real-life photos on the Internet that have been turned into comics. They are very amazing and full of cuteness. So I went to find it and found an open source project that converts photos to comics - AnimeGANv2. There is a project address at the end of the article. I like to remember to bookmark and like it.
Article directory
Introduction
AnimeGANv2 is an open source image-to-manga project developed with python based on tensorflow. It has accumulated to 4K star, which is a very good project.
Effect
The school fences and bridges are obviously animated, with a proper two-dimensional style. Street scenes and pedestrians are also caricatured. Cute and cute cartoon style~
Instructions
necessary conditions
-
python 3.6
-
tensorflow-gpu 1.15.0 (GPU 2080Ti, cuda 10.0.130, cudnn 7.6.0)
-
opencv
-
tqdm
-
numpy
-
glob
-
argparse
-
onnxruntime (If onnx file needs to be run.)
Picture to comic
python test.py --checkpoint_dir checkpoint/generator_Hayao_weight --test_dir dataset/test/HR_photo --save_dir Hayao/HR_photo
After the transfer, the bus became cute and cute.
video to animation
python video2anime.py --video video/input/お花見.mp4 --checkpoint_dir checkpoint/generator_Hayao_weight --output video/output
Videos can also be fully converted to animation.
summary
Is this framework very strong? Videos and pictures can be turned into comic style and used to make comic avatars. Comic videos are still pretty good. If you are interested, please try it out~
project address
https://github.com/TachibanaYoshino/AnimeGANv2