An app using a neural network approach to colorize images and videos.
The application (the neural net rather) "hallucinates" possible colors (so an acceptable result is a probable result, not necessarily the real truth).
GUI written in DearPyGui to allow a pleasant user experience.
- Clone this repository
- Download the trained model from Colorization Release v2 Caffemodel
- For video reproduction, you need MPV: MPV Installation, or quickly for Debian based distros:
sudo apt install mpv
- For displaying images, you need sxiv:
sudo apt install sxiv
- Other packages that you need can be installed with:
sudo apt install python3-tk python3-dev python3-pip scrot
- Execute
pip3 install -r requirements.txt
to download all necessary libraries - Execute with
python3 main.py
or./main.py
- Colorize images
- Colorize videos
- Colorize images & videos at the same time
- Cancel video colorization if it is taking too long for your liking
- Video colorization progress bar
- File picker, no need for program arguments
- RGB information on mouse hover
- Zhang, R., Isola, P., & Efros, A. A. (2016). Colorful Image Colorization. ECCV.
- Zhang, R., Zhu, J.-Y., Isola, P., Geng, X., Lin, A. S., Yu, T., & Efros, A. A. (2017). Real-Time User-Guided Image Colorization with Learned Deep Priors. ACM Transactions on Graphics (TOG), 9(4).
- Rosebrock A. (2019). Black and white image colorization with OpenCV and Deep Learning
- DearPyGui documentation.