You can use it to automatically remove the mosaics in images and videos, or add mosaics to them.
This porject based on "semantic segmentation" and "Image-to-Image Translation".
origin | auto add mosaic | auto clean mosaic |
---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
- Compared with DeepCreamPy
mosaic image | DeepCreamPy | ours |
---|---|---|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
- Style Transfer
origin | to Van Gogh | to winter |
---|---|---|
![]() |
![]() |
![]() |
An interesting example:Ricardo Milos to cat
You can either run DeepMosaics via pre-built binary package or from source.
For windows, we bulid a GUI version for easy test.
Download this version and pre-trained model via [Google Drive] [百度云,提取码1x0a]
- Require Windows_x86_64, Windows10 is better.
- Different pre-trained models are suitable for different effects.[Introduction to pre-trained models]
- Run time depends on computer performance(The current version does not support gpu, if you need to use gpu please run source).
- If output video cannot be played, you can try with potplayer.
- GUI version update slower than source.
- Linux, Mac OS, Windows
- Python 3.6+
- ffmpeg 3.4.6
- Pytorch 1.0+
- CPU or NVIDIA GPU + CUDA CuDNN
This code depends on opencv-python, torchvision available via pip install.
git clone https://github.com/HypoX64/DeepMosaics
cd DeepMosaics
You can download pre_trained models and put them into './pretrained_models'.
[Google Drive] [百度云,提取码1x0a]
[Introduction to pre-trained models]
- Add Mosaic (output media will save in './result')
python3 deepmosaic.py --media_path ./imgs/ruoruo.jpg --model_path ./pretrained_models/mosaic/add_face.pth --use_gpu 0
- Clean Mosaic (output media will save in './result')
python3 deepmosaic.py --media_path ./result/ruoruo_add.jpg --model_path ./pretrained_models/mosaic/clean_face_HD.pth --use_gpu 0
If you want to test other image or video, please refer to this file.
[options_introduction.md]
If you want to train with your own dataset, please refer to training_with_your_own_dataset.md
This code borrows heavily from [pytorch-CycleGAN-and-pix2pix] [Pytorch-UNet] [pix2pixHD] [BiSeNet].