This project aims to produce videos in the style of "picasa face movie". Example (generated by picasa) here: https://www.youtube.com/watch?v=fLQtssJDMMc
-
Create a python environment (or don't, you do you)
python -mvenv env . env/bin/activate
-
Install dependencies
pip install -r requirements.txt
-
Run the gui to select images and faces. When you're done, save your project. It's a json file but the extension doesn't matter.
python gui.py
When you have a project file, you can generate the movie from the command line.
python generate_facemovie.py <project.json> movie.mov
- Work around the possible mismatches between faces detected by
mediapipe
andface_recognition
libraries - GUI: allow adding single image
- GUI: allow removing single image
- GUI: allow reordering images
- Action buttons with icons
- menu bar
- Sometimes the oriented image eyes are not exactly horizontal. Debug and fix this
- Allow picking the eye target location
- Packaging
- Choose a non-generic file extension for project files
- Improve CLI: allow overwriting settings from project json