Simple wrapper UI for OpenSeeFace's facetracker.
- Start / Stop the tracker
- Select Webcam
- Select video mode (width, height, frames per second)
- Select tracking model used by the facetracker
- Set IP and Port for the tracker to listen
- gcc
- cmake
- cairo-devel
- python3-devel
- python3-pip
- gobject-introspection-devel
- v4l-utils
- typelib-1_0-Gtk-4_0
- typelib-1_0-Adw-1
- Install Builder
- Open Source Directory of Facetracker
- Run
- Clone Repository
- Install
- typelib-1_0-Gtk-4_0
- typelib-1_0-Adw-1
- Provide osf pre-build binary (see below)
python3 facetracker/__init__.py
To ease up flatpak distribution Facetracker uses OpenSeeFace's facetracker as a pre-build binary.
Get is as follows:
git clone git@github.com:emilianavt/OpenSeeFace.gitcd OpenSeeFacepip install onnxruntime opencv-python pillow numpypip install -U pyinstallerpyinstaller ./facetracker.pycp models ./dist/facetracker/
To test the pre-build locally create a symbolic link inside the source directory of Facetracker/facetracker/osf pointing towards the dist directory of the pre-build.
The following cams have been tested with this application in functioning and gathering device capabilities properly
- NB Pro: BisonCam
- Logitec C922 Pro Stream Webcam
- USB3.0 capture (yes an actual capture card)
- Logitec HD WebCam C270
- pulsonic HDR webcam
- Integrated Camera of Lenovo V15 G4 AMN